CN106372662A - Helmet wearing detection method and device, camera, and server - Google Patents
Helmet wearing detection method and device, camera, and server Download PDFInfo
- Publication number
- CN106372662A CN106372662A CN201610778655.0A CN201610778655A CN106372662A CN 106372662 A CN106372662 A CN 106372662A CN 201610778655 A CN201610778655 A CN 201610778655A CN 106372662 A CN106372662 A CN 106372662A
- Authority
- CN
- China
- Prior art keywords
- safety helmet
- people
- detection
- training
- video image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a helmet wearing detection method and device, a camera, and a server. The method comprises: obtaining a scene video image, detecting the scene video image to obtain human position through a human position detection model built by training learning, and determining whether the human position is located in a helmet wearing area or not; if the human position is located in the helmet wearing area, detecting the scene video image to obtain the human head helmet combination state through a human helmet combination detection model built by training learning, and determining whether the helmet is worn or not according to the human helmet combination state, if the helmet is not worn, not passing through the wearing verification of the helmet and if the helmet is worn, detecting the scene video image to obtain the type of the helmet through the helmet type detection model built by training learning, determining if the type of the helmet accords with the rule or not, if the type of the helmet accords with the rule, passing through the wearing verification of the helmet, or else, not passing through the wearing verification of the helmet, and therefore, the accuracy of the detection result of the helmet wearing is improved.
Description
Technical field
The present invention relates to field of computer technology, the detection method that more particularly to a kind of safety helmet is worn and device, taking the photograph
As head, server.
Background technology
Safety helmet is the Head protective devices of head impact when anti-object hits and falls, and construction worker passes through safe wearing
Cap, in order to protect head, from the object injury fallen.But it is frequently present of construction worker less than safe wearing cap according to the rules
Situation occur, the wear condition of safety helmet is monitored in real time, and detect whether according to the rules safe wearing cap to close weight
Will.
The detection method that traditional safety helmet is worn, by the rgb component detection of image whether safe wearing cap, by light
The impact of line strength is larger, and for black safety helmet due to close with hair color it is impossible to good detect, and can only identify
Whether according to the rules whether safety helmet worn it is impossible to judge workman's safe wearing cap, leads to the testing result that safety helmet is worn
Accuracy is not high.
Content of the invention
Based on this it is necessary to be directed to above-mentioned technical problem, provide a kind of safe wearing cap according to the rules of judging whether
Detection method and device, photographic head, server that safety helmet is worn, improve the accuracy of the testing result that safety helmet is worn.
The detection method that a kind of safety helmet is worn, methods described includes:
Obtain live video image, the position of human body detection model set up by training study is to described live video image
Carry out detection and obtain position of human body, judge whether to wear region in safety helmet according to described position of human body；
If it is, being entered to described live video image by the number of people safety helmet joint-detection model that training study is set up
Row detection obtains number of people safety helmet united state, judges whether safe wearing cap according to described number of people safety helmet united state, such as
Fruit there is no safe wearing cap, then safety helmet wear not verified；
If safe wearing cap, the safety helmet type detection model set up by training study is to described live video figure
Obtain safety helmet type as carrying out detection, judge whether described safety helmet type meets regulation, if meeting regulation, safety helmet
Wear by checking, otherwise, safety helmet is worn not verified.
The detection means that a kind of safety helmet is worn, described device includes:
Acquisition module, for obtaining live video image；
Region detection module worn by safety helmet, by the position of human body detection model that training study is set up, described scene is regarded
Frequency image carries out detection and obtains position of human body, judges whether to wear region in safety helmet according to described position of human body, if it is,
Enter number of people safety helmet united state detection module；
Number of people safety helmet united state detection module, for the number of people safety helmet joint-detection mould set up by training study
Type carries out detection and obtains number of people safety helmet united state to described live video image, according to described number of people safety helmet united state
Judge whether safe wearing cap, without safe wearing cap, then safety helmet wear not verified, if safe wearing cap,
Then enter safety helmet type detection module；
Safety helmet type detection module, for the safety helmet type detection model by training study foundation to described scene
Video image carries out detection and obtains safety helmet type, judges whether described safety helmet type meets regulation, if meeting regulation,
Safety helmet is worn by checking, and otherwise, safety helmet is worn not verified.
A kind of photographic head, the detection means worn including the safety helmet described in any of the above-described embodiment.
A kind of server is it is characterised in that what safety helmet that described server is included described in any of the above-described embodiment was worn
Detection means.
Detection method and device, photographic head, server that above-mentioned safety helmet is worn, by obtaining live video image, lead to
Cross the position of human body detection model that training study sets up to carry out detection to live video image and obtain position of human body, according to people's position
Put and judge whether to wear region in safety helmet, if it is, the number of people safety helmet joint-detection model set up by training study
Live video image is carried out with detection and obtains number of people safety helmet united state, judge whether to wear according to number of people safety helmet united state
Wear a safety helmet, without safe wearing cap, then safety helmet wear not verified, if safe wearing cap, by training
The safety helmet type detection model that study is set up carries out detection to live video image and obtains safety helmet type, judges safety helmet class
Whether type meets regulation, if meeting regulation, safety helmet is worn by checking, and otherwise, safety helmet is worn not verified.?
During detection is worn to safety helmet, using the position of human body detection model set up by training study, number of people safety helmet connection
Close detection model, safety helmet type detection model detects to live video image, respectively because each detection model is
It is trained obtaining by substantial amounts of real scene training data, the accuracy of testing result is high, and by detecting layer by layer, not only
Ensure that safe wearing cap, it is ensured that the style complies with set of safe wearing cap, improves safety helmet and wears the credible of testing result
Degree and accurateness, strengthen the intensity of security protection.
Brief description
Fig. 1 is the applied environment figure of the detection method that safety helmet is worn in an embodiment；
Fig. 2 is the cut-away view of photographic head in Fig. 1 in an embodiment；
Fig. 3 is the cut-away view of server in Fig. 1 in an embodiment；
Fig. 4 is the flow chart of the detection method that safety helmet is worn in an embodiment；
Fig. 5 is the flow chart setting up number of people safety helmet joint-detection model in an embodiment；
Fig. 6 is the flow chart determining number of people safety helmet united state in an embodiment；
Fig. 7 is to set up safety helmet detection model, position of human body detection model and number of people detection model in an embodiment
Flow chart；
Fig. 8 is the detailed process schematic diagram of the detection method that safety helmet is worn in a specific embodiment；
Fig. 9 is the structured flowchart of the detection means that safety helmet is worn in an embodiment；
Figure 10 is the structured flowchart of the detection means that safety helmet is worn in another embodiment；
Figure 11 is the structured flowchart of number of people safety helmet united state detection module in an embodiment；
Figure 12 is the structured flowchart of the detection means that safety helmet is worn in further embodiment；
Figure 13 is the structured flowchart of the detection means that safety helmet is worn in another embodiment；
Figure 14 is the structured flowchart of the detection means that safety helmet is worn in another embodiment；
Figure 15 is the structured flowchart of the detection means that safety helmet is worn in further embodiment.
Specific embodiment
Fig. 1 is the applied environment figure that the detection method that in an embodiment, safety helmet is worn is run.As shown in figure 1, should
Include terminal 110, photographic head 120 server 130 with environment, wherein terminal 110, photographic head 120 server 130 passes through net
Network is communicated.It is understood that increasing can be carried out to the equipment in above-mentioned applied environment and adds deduct according to the situation of deployment
Few.
Terminal 110 can be smart mobile phone, panel computer, notebook computer, desk computer etc., but is not limited thereto.
Server 130 can be individual server or server cluster, and photographic head 120 can be one or more, by inputting photographic head 120
The live video image of collection is worn to safety helmet to the detection model by training study foundation and is detected.Can be by photographic head
The live video image of 120 collections is transferred to server 130 concentration and is detected, server 130 can be the meter of building site deployment
Calculation machine/server or public cloud/privately owned Cloud Server, or in photographic head front end installation detecting device, transported by detection means
Row video analysis algorithm is detected, live video image is carried out position of human body detection, the number of people safety helmet united state detection,
A series of detection process such as safety helmet type detection, obtains safety helmet after comprehensive analysis and wears testing result, and can basis
Testing result sends prompting message to terminal 110, so as real time inspection testing result with take safety precautions in time.
In one embodiment, in Fig. 1 photographic head 120 internal structure as shown in Fig. 2 include input equipment, processor,
Storage medium, internal memory, interface, input equipment is used for gathering video image, and processor calculates and control ability for providing,
By detection model, video image is tested and analyzed, support the operation of whole photographic head 120, storage medium is stored with first
The detection means that safety helmet is worn, this device is used for realizing being applied to the detection method that a kind of safety helmet of photographic head 120 is worn.
The operation inside saving as the detection means that the first safety helmet in storage medium is worn provides environment, and interface is used for carrying out with server
Data transfer, interface can be data line interface or wave point etc..
In one embodiment, the internal structure of the server 130 in Fig. 1 is as shown in figure 3, this server 130 includes leading to
Cross processor, storage medium, internal memory and the network interface of system bus connection.Wherein, the storage medium storage of this server 130
There is the detection means that operating system, data base and the second safety helmet are worn, data base is used for data storage, such as storage live video
Image, the detection means that the second safety helmet is worn is used for realizing the detection side that a kind of safety helmet being applied to server 130 is worn
Method.The processor of this server 130 is used for providing calculating and control ability, supports the operation of whole server 130.This server
The operation saving as the detection means that the second safety helmet in storage medium is worn in 130 provides environment.The net of this server 130
Network interface is used for communicating by network connection with outside terminal 110, photographic head 120, such as sends prompting message to terminal 110
Deng.
In one embodiment, there is provided the detection method that a kind of safety helmet is worn, to be applied in above-mentioned applied environment
Photographic head or server illustrating, comprise the following steps:
Step s210, obtains live video image, by the position of human body detection model that training study is set up, scene is regarded
Frequency image carries out detection and obtains position of human body, judges whether to wear region in safety helmet according to position of human body, if it is, entering
Step s220.
Specifically, can the key position deployment photographic head such as high point in construction site, gateway of lift, gather work
Ground live video image.If the detection method that safety helmet is worn is applied to photographic head, what direct access photographic head gathered shows
Live video image, if being applied to server, can be sent and be processed to server by field video image, the scene when sending
Video image carries corresponding camera identification, is easy to distinguish the collection position of live video image, and subsequently through photographic head
Mark carries out safety helmet type and the selection in region worn by safety helmet.Because photographic head is the video of continuous acquisition, can lead to
Cross the frequency of the video image that custom algorithm determination is detected, detect once, then according to frequency from continuous acquisition within such as 5 seconds
Extract corresponding frame of video in video, then detected, reduce the frame number of detection, to reduce complexity.Also can be known by video
Corresponding video frequency, in the case of moving object has been detected, is just carried out safety helmet and wears detection, keep away further by other algorithm
Exempt from current scene nobody, the static frames only having powerful connections take a large amount of detection resources, it is to avoid invalid detection.Learnt by training
The position of human body detection model set up carries out detection and obtains position of human body to live video image, and the training pattern of employing can be
One or more algorithms concrete in machine learning, deep learning, such as support vector machine, adaboosting, convolutional Neural
Network etc..The detection model obtaining is exactly specific algorithm according to the result after training data study.Off-line training can be passed through
Mode obtains accurate model using substantial amounts of training data.For position of human body detection model, for the positive sample containing human body
This picture can build the primary vector related to position of human body, and the negative master drawing piece not containing human body can build secondary vector, will not
Same picture and corresponding vector input after training pattern is trained and obtain training pattern unknown parameter so that it is determined that obtaining people
Body position detection model.After obtaining position of human body detection model, then input test picture, just can automatically calculate corresponding human body
Position.Because early stage establishes model by substantial amounts of true photographed scene picture so that the output result accuracy of model is high.
In one embodiment, for different photographic head, obtain the corresponding training data of each camera identification, according to
The scene activity characteristics of image that photographic head shoots adopts different position of human body to detect training pattern, obtains photographic head through training
Identify corresponding different position of human body detection model, because the structure of position of human body detection model considers photographic head shooting
Scene activity feature, the picture background that same photographic head shoots is identical, can according to the complexity of background picture and this
The state of the moving object of scape is using the position of human body detection training pattern being suitable for, thus improving position of human body detection mould further
The accuracy of type.Scene as the first photographic head shoots is the first building site, and the scene that second camera shoots is the second building site, by
Different with the task in the second building site in the first building site, lead to the human body exercise amount in the second building site to be more than the human body in the first building site
Quantity of motion, the big I of human body exercise amount is weighed by the size of the motion vector of video image and is quantified, thus according to fortune
The size of momentum builds suitable position of human body detection model.Also can be by the position of human body detection model building according to the field being suitable for
Scape is classified, and when there being new photographic head to dispose, can determine its corresponding scene by the feature of the video image of its collection,
Thus according to the corresponding relation of scene and position of human body detection model, determining target body position detection model, thus new having
Photographic head deployment when, position of human body detection model can be obtained in real time, without obtaining people again by way of off-line training
Body position detection model, quick and convenient.
Because the scene that each photographic head shoots is different, it is also different that area worn by the corresponding safety helmet of each scene, can be in advance
The corresponding safety helmet of each scene is set and wears area, the region of safe wearing cap manually can be indicated in advance, such as in video
In picture, respective regions draw rectangle frame, polygon is superimposed upon in current collection video image, are easy to observe, according to position of human body
Judge whether that area worn by the safety helmet in current scene.The panoramic range of the scene shooting for current photographic head is all safety helmet
When wearing area, this step can directly be skipped.Also rationally photographic head can be disposed so that area worn by the corresponding safety helmet of each photographic head
Fixing with respect to the position of current scene, thus needing not distinguish between each photographic head, simplifying the peace judging whether in current scene
The step that area worn by full cap.
Step s220, is examined to live video image by the number of people safety helmet joint-detection model that training study is set up
Record number of people safety helmet united state, safe wearing cap is judged whether according to number of people safety helmet united state, without pendant
Wear a safety helmet, then safety helmet wear not verified, if safe wearing cap, enter step s230.
Specifically, number of people safety helmet joint-detection model is used for detecting the number of people whether safe wearing cap, can be single model,
Directly input test pictures and just can directly obtain number of people safety helmet united state to a model.It is alternatively multiple submodels to join
Conjunction respectively obtains multiple testing results, determines number of people safety helmet united state further according to the relation between multiple testing results.People
Head safety helmet united state is divided into normal condition and abnormality, and normal condition refers to safe wearing cap, and abnormality refers to not
Safe wearing cap.
Detection is carried out to live video image by the number of people safety helmet joint-detection model that training study is set up and obtains people
Head safety helmet united state, the training pattern of employing can be machine learning, one or more calculations concrete in deep learning
Method, such as support vector machine, adaboosting, convolutional neural networks etc..The detection model obtaining be exactly specific algorithm according to
Result after training data study.By way of off-line training, accurate model can be obtained using substantial amounts of training data.Right
In number of people safety helmet joint-detection model, if single model, then the positive sample picture for safe wearing cap marks the number of people
Safety helmet joint normal condition, for the negative master drawing piece mark number of people safety helmet joint abnormality of non-safe wearing cap, will not
With picture and corresponding state constitute and obtain training pattern unknown parameter after vector input training pattern is trained thus really
Surely obtain number of people safety helmet joint-detection model.After obtaining number of people safety helmet joint-detection model, then input test picture, with regard to energy
Automatically calculate corresponding number of people safety helmet united state.If number of people safety helmet joint-detection model includes multiple submodels,
Then obtain the training data of each submodel respectively and corresponding markup information forms sub- training data, by each sub- training data
Input respectively after corresponding sub- training pattern is trained and obtain unknown parameter so that it is determined that obtaining each submodel.Calculate the number of people
During safety helmet united state, test pictures are inputted each submodel, respectively obtain the corresponding testing result of submodel, then basis
Relation between each testing result is calculated number of people safety helmet united state.Between wherein how according to each testing result
Relation be calculated the specific algorithm of number of people safety helmet united state can be self-defined as needed.The number of people peace that terminal obtains
Full cap united state if exception, does not then have safe wearing cap, safety helmet is worn not verified.If normal, then wear
Wear a safety helmet, enter the detection of next step.
Step s230, carries out detecting to live video image by the safety helmet type detection model that training study is set up
To safety helmet type, judge whether safety helmet type meets regulation, if meeting regulation, safety helmet is worn by checking, no
Then, safety helmet wear not verified.
Specifically, different building sites and task, due to the difference of degree of danger, need to wear different types of safety
Cap.Different types of safety helmet has different shapes, color etc..The number of people safety helmet type detection set up by training study
Model carries out detection and obtains safety helmet type to live video image, and the training pattern of employing can be machine learning, depth
One or more algorithms concrete in habit, such as support vector machine, adaboosting, convolutional neural networks etc..The inspection obtaining
Surveying model is exactly specific algorithm according to the result after training data study.Substantial amounts of instruction can be adopted by way of off-line training
Practice data and obtain accurate model.For safety helmet type detection model, then the type for each safety helmet in picture is carried out
Mark, the type of different pictures and corresponding safety helmet is constituted after vector input training pattern is trained and obtains training mould
Type unknown parameter is so that it is determined that obtain safety helmet type detection model.After obtaining safety helmet type detection model, then input test
Picture, just can calculate the type of corresponding safety helmet automatically.Judge the type whether standard security cap with regulation of safety helmet
Type matching, if coupling; would meet regulation, otherwise against regulation.For different scene corresponding standard security caps
When different, need first to obtain currently affiliated scene, then obtain current scene corresponding standard security cap type, then judge to examine
Whether corresponding with the current scene standard security cap type matching of the safety helmet type measured, wherein standard security cap type is permissible
It is the level of security scope of safety helmet, or Type Range.Currently affiliated scene can be by the picture feature analysis of current collection
Obtain, corresponding relation that also can directly according to camera identification and scene, first obtain camera identification, then obtain corresponding current
Scene.Need safe wearing cap not to the utmost in addition it is also necessary to the type of safety helmet correctly just calculates safety helmet wears by checking so that safety
The result that detection worn by cap is more accurate, strengthens the intensity of security protection.
In the present embodiment, by obtaining live video image, by the position of human body detection model pair of training study foundation
Live video image carries out detection and obtains position of human body, judges whether to wear region in safety helmet according to position of human body, if it is,
Then detection is carried out to live video image by the number of people safety helmet joint-detection model that training study is set up and obtain number of people safety
Cap united state, judges whether safe wearing cap according to number of people safety helmet united state, without safe wearing cap, then safety
Cap is worn not verified, if safe wearing cap, by the safety helmet type detection model of training study foundation to scene
Video image carries out detection and obtains safety helmet type, judges whether safety helmet type meets regulation, if meeting regulation, safety
Cap is worn by checking, and otherwise, safety helmet is worn not verified.During detection is worn to safety helmet, using passing through
The position of human body detection model of training study foundation, number of people safety helmet joint-detection model, safety helmet type detection model are respectively
Live video image is detected, because each detection model is to be trained by substantial amounts of real scene training data
Obtain, the accuracy of testing result is high, and by detecting layer by layer, not only ensure safe wearing cap it is ensured that safe wearing cap
Style complies with set, improves credibility and the accurateness that testing result worn by safety helmet, strengthens the intensity of security protection.
In one embodiment, number of people safety helmet joint-detection model be single model, as shown in figure 5, step s210 it
Before, also include:
Step s310, obtains training video image, and training video image includes the positive sample image of human body safe wearing cap
Negative sample image with human body not safe wearing cap.
Step s320, obtains the markup information of training video image, the markup information of positive sample image is number of people safety helmet
United state is normal, and the markup information of negative sample image is number of people safety helmet united state exception, training video image and correspondence
Markup information composition training data.
Step s330, training data is inputted number of people safety helmet joint-detection training pattern, the method based on training study
Training data is trained obtain number of people safety helmet joint-detection model.
Specifically, human body safe wearing cap is positive sample image, and markup information is that number of people safety helmet united state is normal, people
Safe wearing cap is not negative sample image to body, and markup information is that number of people safety helmet united state is abnormal, and the number of people and safety helmet are made
Obtain training data for an entirety thus setting up model it is only necessary to one model of input just can obtain number of people safety helmet joint shape
State is it is not necessary to analyze number of people position and safety helmet position respectively, simple and convenient.
In one embodiment, number of people safety helmet joint-detection model includes number of people detection model and the detection of safety helmet position
Model, as shown in fig. 6, pass through the number of people safety helmet joint-detection model of training study foundation to live video figure in step s220
As carrying out detecting that the step obtaining number of people safety helmet united state includes:
Step s221, carries out detection according to number of people detection model to live video image and obtains people's head region.
Specifically, number of people detection model is the detection model set up by way of off-line training learns, for scene
Video image carries out detection and obtains people's head region, and the training pattern of employing can be machine learning, in deep learning concrete one
Plant or many algorithms, such as support vector machine, adaboosting, convolutional neural networks etc., accuracy is high.Input scene regards
Frequency image can directly obtain the people's head region detecting, can be according to mark rectangle in the scope of people's head region at the scene video image
Frame, irregular shape frame etc..
Step s222, carries out detection according to safety helmet position detection model to live video image and obtains safety helmet region.
Specifically, safety helmet position detection model is the detection model set up by way of off-line training learns, and is used for
Live video image is carried out with detection and obtains safety helmet region, the training pattern of employing can be machine learning, in deep learning
One or more algorithms concrete, such as support vector machine, adaboosting, convolutional neural networks etc., accuracy is high.Defeated
Enter live video image and can directly obtain the safety helmet region detecting, can be according to the scope in safety helmet region video figure at the scene
As upper mark rectangle frame, irregular shape frame etc..
Step s223, the distance according to people's head region and safety helmet region and overlapping area obtain number of people safety helmet joint shape
State.
Specifically, the computational methods of the distance in people's head region and safety helmet region can be self-defined as needed, can pass through two
The range difference of individual Region Matching location point obtains, such as the line segment with number of people regional center point and safety helmet regional center point as end points
Corresponding length is distance, or between number of people regional center point and safety helmet regional center point coordinate points in the horizontal direction away from
With a distance from from length being, or the distance between the point in people's head region and safety helmet region upper left corner coordinate points in the horizontal direction length
It is distance.Area according to people's head region and safety helmet region and overlapping area calculate number of people safety helmet joint proportionality coefficient, such as
Fruit number of people safety helmet joint proportionality coefficient is more than first threshold, and the distance in people's head region and safety helmet region is less than the first threshold
Value, then number of people safety helmet united state is normal, and otherwise number of people safety helmet united state is abnormal.
In one embodiment, between number of people regional center point and safety helmet regional center point coordinate points in the horizontal direction
Be d apart from length, the area of people's head region is s1, and the area in safety helmet region is s2, and overlapping area is s0, number of people safety helmet
Joint proportionality coefficient e=s0/ (s1+s2-s0).
In the present embodiment, number of people safety helmet is calculated by number of people detection model and the cooperation of safety helmet position detection model
United state, can customize the distance according to people's head region and safety helmet region during calculating and overlapping area obtains people
During head safety helmet united state, number of people safety helmet united state is normal condition, and design conditions can customize, flexibly and easily.
In one embodiment, as shown in fig. 7, before step s210, also including:
Step s410, obtain training video image, training video image include the positive sample image containing examined object,
Do not contain the negative sample image of examined object, examined object is human body, the number of people, at least one in safety helmet.
Step s420, obtains the markup information set of training video image, and markup information set includes human region mark
Information, number of people area marking information and safety helmet area marking information.
Specifically, the image of photographic head collection does not have markup information, first can send the image that photographic head gathers to end
End, is manually marked the markup information obtaining each image, markup information intuitively can be marked by way of figure, such as square
The form mark of shape frame, such as human body markup information is exactly the minimum rectangle frame containing human body.
Step s430, obtains the corresponding current markup information of current training pattern from markup information set, current mark letter
Breath forms current training data with training video image.
Specifically, if current training pattern is safety helmet position detection training pattern, obtaining current markup information is
Each training video image corresponding safety helmet area marking information, and correspond with training video image, each training regards
Frequency image and corresponding safety helmet area marking information composition current safety cap detection training data.
Step s440, current training data is inputted current training pattern, and the method based on training study is trained to current
Data is trained obtaining corresponding current detection model, and current training pattern includes safety helmet position detection training pattern, people
Body position detection training pattern and number of people detection training pattern, current detection model includes described safety helmet detection model, human body
Position detection model and number of people detection model.
Specifically, using methods such as machine learning, deep learnings, align sample image and the training of negative sample image is worked as
Front detection model is it can be determined that go out the particular location whether a picture includes target object and target object.As safety helmet
Detection model can determine whether whether picture to be detected includes the particular location of safety helmet and safety helmet.Pacified by off-line training
After full cap detection model, position of human body detection model and number of people detection model, altimetric image to be checked can be directly inputted online and pacified
Full cap position, position of human body and number of people position.
In one embodiment, method is further comprising the steps of: the safety helmet state-detection mould set up using training study
Type carries out detection to live video image and obtains safety helmet current state, judges whether safety helmet damages according to safety helmet current state
Bad, if safety helmet damages, safety helmet is worn not verified.
Specifically, during the safety helmet state-detection model of foundation, the training pattern of employing can be machine learning, depth
One or more algorithms concrete in habit, such as support vector machine, adaboosting, convolutional neural networks etc..The inspection obtaining
Surveying model is exactly specific algorithm according to the result after training data study.Substantial amounts of instruction can be adopted by way of off-line training
Practice data and obtain accurate model.For safety helmet state-detection model, prepare a collection of different safety helmet photo, according to this peace
, if appropriate for being continuing with manually being marked state, if can be continuing with, safety helmet state is normal for full cap, otherwise, safety
Cap abnormal state.Training pattern is obtained unknown after different photos and corresponding labeled data input training pattern are trained
Parameter is so that it is determined that obtain safety helmet state-detection model.Safety helmet state-detection model is a sorter model, can be right
Safety helmet photo to be detected provides safety helmet state-detection as a result, it is possible to represent safety by the form of fraction or judgment value
Hat shape state testing result.If safety helmet current state represents that safety helmet damages, immediately arrive at safety helmet and wear and do not pass through to test
Card.The testing process that safety helmet is worn can be interrupted in advance, safety helmet state normally safety helmet wear by checking before
Put forward condition, further increase correctness and the reliability that testing result worn by safety helmet, strengthen the intensity of safeguard protection.
In one embodiment, method also includes: obtains the corresponding camera identification of live video image.In step s210
Live video image is carried out detect by the position of human body detection model that training study is set up obtain position of human body step it
Before, also include: judge whether the corresponding shooting area of camera identification is that region worn by full view safety cap, if it is, judging
Wear region in safety helmet, be directly entered step s220.Otherwise enter back into the position of human body by training study is set up and detect mould
Type carries out to described live video image detecting the step obtaining position of human body.
Specifically, the corresponding relation that camera identification and safety helmet are worn between region can be stored in advance, thus judging to take the photograph
As leader knows whether corresponding shooting area is that region worn by full view safety cap, if it is, human body is any in photo current
Position is all to wear region in safety helmet, can be directly entered step s220, improves the efficiency that detection worn by safety helmet.
Judge whether that the step wearing region in safety helmet includes according to position of human body in step s210: according to shooting leader
Know and obtain corresponding current region safety helmet type, if safety helmet type and current region safety helmet type
Join, then safety helmet style complies with set, otherwise safety helmet type mismatch closes regulation.
Specifically, the corresponding relation of camera identification and current region safety helmet type can be stored in advance, different
Camera identification represents different building site regions, due to the difference of any and scene of working, generally requires to wear dissimilar
Safety helmet, according to camera identification obtain current region safety helmet type so that the detection that safety helmet is worn more meets
The demand of different scenes, further increases reliability and the safety of the detection that safety helmet is worn.
In one embodiment, after not verified step worn by safety helmet, also include: send safety helmet and wear not
By the reminder message of checking to terminal, reminder message includes video, image, at least one in voice, and reminder message carries
The reason safety helmet wears not verified information.
Specifically, if safety helmet wear not verified in time management personnel can be notified by reminder message, can will carry
Awake message sends to the hand held equipment terminal of management personnel, also can be reported by the form of broadcast and alert.Reminder message
Can be at least one in video, image, voice.In the image of output, safety helmet region, people can be indicated by wire frame
Body position region and the number of people band of position, so that management personnel's quick recognition detection result.
In a specific embodiment, the detailed process of the detection method that safety helmet is worn is as shown in Figure 8, comprising:
Input live video image, carries out human detection by position of human body detection model respectively, detects mould by the number of people
Type carries out number of people detection, carries out safety helmet detection by safety helmet position detection model, wherein obtains people's position after human detection
Put, judge whether in the region needing safe wearing cap, after safety helmet detection, obtain the position of safety helmet, in conjunction with number of people detection
The number of people position judgment obtaining whether safe wearing cap.Again safety helmet type detection is carried out by safety helmet type detection model,
Obtaining the type of safety helmet, if having worn safety helmet, judging whether to wear the safety helmet of stated type, finally by safety
The detection of hat shape state detection model obtains safety helmet state, judges whether to need to change safety helmet, if not needing to change, safety
The testing result that cap is worn is that otherwise testing result is not verified by checking.
In one embodiment, as shown in Figure 9, there is provided the detection means that a kind of safety helmet is worn, comprising:
Acquisition module 510, for obtaining live video image.
Region detection module 520 worn by safety helmet, by the position of human body detection model that training study is set up, scene is regarded
Frequency image carries out detection and obtains position of human body, judges whether to wear region in safety helmet according to position of human body, if it is, entering
Number of people safety helmet united state detection module.
Number of people safety helmet united state detection module 530, for combining inspection by the number of people safety helmet that training study is set up
Survey model carries out detection and obtains number of people safety helmet united state to live video image, is judged according to number of people safety helmet united state
Whether safe wearing cap, without safe wearing cap, then safety helmet is worn not verified, if safe wearing cap, enters
Enter safety helmet type detection module.
Safety helmet type detection module 540, for the safety helmet type detection model by training study foundation to scene
Video image carries out detection and obtains safety helmet type, judges whether safety helmet type meets regulation, if meeting regulation, safety
Cap is worn by checking, and otherwise, safety helmet is worn not verified.
In one embodiment, number of people safety helmet joint-detection model is single model, and as shown in Figure 10, device also wraps
Include:
Number of people safety helmet joint-detection model building module 550, for obtaining training video image, training video image bag
Include the negative sample image of the positive sample image of human body safe wearing cap and human body not safe wearing cap, obtain training video image
Markup information, the markup information of positive sample image is that number of people safety helmet united state is normal, and the markup information of negative sample image is
Number of people safety helmet united state is abnormal, and training video image and corresponding markup information form training data, and training data is defeated
Enter number of people safety helmet joint-detection training pattern, training data is trained obtain number of people safety based on the method for training study
Cap joint-detection model.
In one embodiment, number of people safety helmet joint-detection model includes number of people detection model and the detection of safety helmet position
Model, as shown in figure 11, number of people safety helmet united state detection module 530 includes:
Number of people region detection unit 531, obtains people for carrying out detection according to number of people detection model to live video image
Head region.
Safety helmet region detection unit 532, carries out detecting to live video image according to safety helmet position detection model
To safety helmet region.
Number of people safety helmet united state determining unit 533, for the distance according to people's head region and safety helmet region and weight
Folded area obtains number of people safety helmet united state.
In one embodiment, as shown in figure 12, device also includes:
Position model sets up module 560, and for obtaining training video image, training video image is included containing thing to be detected
The positive sample image of body, do not contain the negative sample image of examined object, examined object is human body, in the number of people, safety helmet
At least one, obtains the markup information set of training video image, and markup information set includes human region markup information, the number of people
Area marking information and safety helmet area marking information, obtain the corresponding current mark of current training pattern from markup information set
Information, current markup information forms current training data with training video image, trains mould by current for the input of current training data
Type, is trained to current training data obtaining corresponding current detection model based on the method for training study, currently trains mould
Type includes safety helmet position detection training pattern, position of human body detection training pattern and number of people detection training pattern, current detection
Model includes described safety helmet detection model, position of human body detection model and number of people detection model.
In one embodiment, as shown in figure 13, device also includes:
Safety helmet state detection module 570, for the safety helmet state-detection model using training study foundation to scene
Video image carries out detection and obtains safety helmet current state, judges whether safety helmet damages according to safety helmet current state, if
Safety helmet damage, then safety helmet wear not verified.
In one embodiment, as shown in figure 14, acquisition module is additionally operable to obtain the corresponding photographic head of live video image
Mark, device also includes:
The direct judge module in region 580 worn by safety helmet, for judging that whether the corresponding shooting area of camera identification be
Region worn by full view safety cap, if it is, judging to wear region in safety helmet, is directly entered the inspection of number of people safety helmet united state
Survey module 530, otherwise enter safety helmet and wear region detection module 520.
Safety helmet is worn region detection module 520 and is additionally operable to be worn according to the camera identification corresponding current safety cap of acquisition
Region, judges whether position of human body wears region in current safety cap.
Safety helmet type detection module 540 is additionally operable to obtain corresponding current region safety helmet standard according to camera identification
Type, if safety helmet type is mated with current region safety helmet type, safety helmet style complies with set, otherwise safety
Cap type mismatch closes regulation.
In one embodiment, as shown in figure 15, device also includes:
Prompting module 590, wears not verified reminder message to terminal for sending safety helmet, reminder message includes
At least one in video, image, voice, information the reason reminder message cap safe to carry wears not verified.
In one embodiment, there is provided a kind of photographic head, photographic head includes the safety helmet described in any of the above-described embodiment
The detection means worn.
Specifically, the detection means that safety helmet is worn is integrated in photographic head, directly safety helmet can be worn and examine
Survey the miscellaneous equipment it is not necessary to extra, simple and convenient.
In one embodiment, there is provided a kind of server includes the inspection that the safety helmet described in any of the above-described embodiment is worn
Survey device.
Specifically, server includes the detection means that safety helmet is worn, and can receive showing of one or more photographic head collection
Video flowing, carries out unifying detection process so that the detection that safety helmet is worn is easy to unified management it is not necessary in each photographic head
Upper installation detecting device, simple and convenient.
One of ordinary skill in the art will appreciate that realizing all or part of flow process in above-described embodiment method, it is permissible
Instruct related hardware to complete by computer program, described program can be stored in a computer read/write memory medium
In, such as in the embodiment of the present invention, this program can be stored in the storage medium of computer system, and by this computer system
At least one computing device, to realize including the flow process of the embodiment as above-mentioned each method.Wherein, described storage medium can be
Magnetic disc, CD, read-only memory (read-only memory, rom) or random access memory (random access
Memory, ram) etc..
Each technical characteristic of embodiment described above can arbitrarily be combined, for making description succinct, not to above-mentioned reality
The all possible combination of each technical characteristic applied in example is all described, as long as however, the combination of these technical characteristics is not deposited
In contradiction, all it is considered to be the scope of this specification record.
Embodiment described above only have expressed the several embodiments of the present invention, and its description is more concrete and detailed, but simultaneously
Can not therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
Say, without departing from the inventive concept of the premise, some deformation can also be made and improve, these broadly fall into the protection of the present invention
Scope.Therefore, the protection domain of patent of the present invention should be defined by claims.
Claims (16)
1. the detection method that a kind of safety helmet is worn, methods described includes:
Obtain live video image, by the position of human body detection model that training study is set up, described live video image is carried out
Detection obtains position of human body, judges whether to wear region in safety helmet according to described position of human body；
If it is, being examined to described live video image by the number of people safety helmet joint-detection model that training study is set up
Recording number of people safety helmet united state, safe wearing cap being judged whether according to described number of people safety helmet united state, if do not had
Have safe wearing cap, then safety helmet wear not verified；
If safe wearing cap, by the safety helmet type detection model that training study is set up, described live video image is entered
Row detection obtains safety helmet type, judges whether described safety helmet type meets regulation, if meeting regulation, safety helmet is worn
By checking, otherwise, safety helmet is worn not verified.
2. method according to claim 1 is it is characterised in that described number of people safety helmet joint-detection model is single mould
Type, described acquisition live video image, the position of human body detection model set up by training study is to described live video image
Before carrying out detecting the step obtaining position of human body, also include:
Obtain training video image, described training video image includes the positive sample image of human body safe wearing cap and human body is not worn
The negative sample image wearing a safety helmet；
Obtain the markup information of described training video image, the markup information of described positive sample image is number of people safety helmet joint shape
State is normal, and the markup information of described negative sample image is that number of people safety helmet united state is abnormal, described training video image and right
The markup information composition training data answered；
Described training data is inputted number of people safety helmet joint-detection training pattern, the method based on training study is to described training
Data is trained obtaining described number of people safety helmet joint-detection model.
3. method according to claim 1 is it is characterised in that described number of people safety helmet joint-detection model includes number of people inspection
Survey model and safety helmet position detection model, the described number of people safety helmet joint-detection model set up by training study is to described
Live video image carries out detecting that the step obtaining number of people safety helmet united state includes:
Detection is carried out to described live video image according to described number of people detection model and obtains people's head region；
Detection is carried out to described live video image according to described safety helmet position detection model and obtains safety helmet region；
Distance according to described people's head region and safety helmet region and overlapping area obtain number of people safety helmet united state.
4. method according to claim 3 is it is characterised in that before the step of described acquisition live video image, also wrap
Include:
Obtain training video image, described training video image includes the positive sample image containing examined object, without needing
The negative sample image of detection object, described examined object is human body, the number of people, at least one in safety helmet；
Obtain the markup information set of described training video image, described markup information set include human region markup information,
Number of people area marking information and safety helmet area marking information；
Obtain the corresponding current markup information of current training pattern from described markup information set, current markup information and training regard
Frequency image forms current training data；
Described current training data is inputted current training pattern, based on the method for training study, described current training data is entered
Row training obtains corresponding current detection model, and described current training pattern includes safety helmet position detection training pattern, human body
Position detection training pattern and number of people detection training pattern, described current detection model includes described safety helmet detection model, people
Body position detection model and number of people detection model.
5. method according to claim 1 is it is characterised in that methods described also includes:
Detection is carried out to described live video image using the safety helmet state-detection model that training study is set up and obtains safety helmet
According to described safety helmet current state, current state, judges whether safety helmet damages；
If safety helmet damages, safety helmet is worn not verified.
6. method according to claim 1 is it is characterised in that methods described also includes:
Obtain the corresponding camera identification of described live video image；
The described position of human body detection model by training study foundation carries out detection to described live video image and obtains human body
Before the step of position, also include:
Judge whether the corresponding shooting area of described camera identification is that region worn by full view safety cap, if it is, judging
Region worn by safety helmet, is directly entered the described number of people safety helmet joint-detection model set up by training study to described scene
Video image carries out detecting the step obtaining number of people safety helmet united state, otherwise enters the described people setting up by training study
Body position detection model carries out to described live video image detecting the step obtaining position of human body；
Judge whether that the step wearing region in safety helmet includes according to described position of human body: obtain according to described camera identification
Region worn by corresponding current safety cap, judges whether described position of human body wears region in described current safety cap；
The described step judging whether described safety helmet type meets regulation includes: is obtained corresponding according to described camera identification
Current region safety helmet type, if described safety helmet type is mated with described current region safety helmet type,
Described safety helmet style complies with set, otherwise described safety helmet type mismatch conjunction regulation.
7. according to claim 1 or 5 method it is characterised in that described safety helmet wear not verified step it
Afterwards, also include:
Send safety helmet and wear not verified reminder message to terminal, described reminder message include video, image, in voice
At least one, information the reason described reminder message cap safe to carry wears not verified.
8. the detection means that a kind of safety helmet is worn is it is characterised in that described device includes:
Acquisition module, for obtaining live video image；
Region detection module worn by safety helmet, and the position of human body detection model set up by training study is to described live video figure
Obtain position of human body as carrying out detection, judge whether to wear region in safety helmet according to described position of human body, if it is, entering
Number of people safety helmet united state detection module；
Number of people safety helmet united state detection module, for the number of people safety helmet joint-detection model pair set up by training study
Described live video image carries out detection and obtains number of people safety helmet united state, is judged according to described number of people safety helmet united state
Whether safe wearing cap, without safe wearing cap, then safety helmet is worn not verified, if safe wearing cap, enters
Enter safety helmet type detection module；
Safety helmet type detection module, for the safety helmet type detection model by training study foundation to described live video
Image carries out detection and obtains safety helmet type, judges whether described safety helmet type meets regulation, if meeting regulation, safety
Cap is worn by checking, and otherwise, safety helmet is worn not verified.
9. device according to claim 8 is it is characterised in that described number of people safety helmet joint-detection model is single mould
Type, described device also includes:
Number of people safety helmet joint-detection model building module, for obtaining training video image, described training video image includes
The negative sample image of the positive sample image of human body safe wearing cap and human body not safe wearing cap, obtains described training video image
Markup information, the markup information of described positive sample image is that number of people safety helmet united state is normal, described negative sample image
Markup information is that number of people safety helmet united state is abnormal, and described training video image and corresponding markup information composition train number
According to by described training data input number of people safety helmet joint-detection training pattern, the method based on training study is to described training
Data is trained obtaining described number of people safety helmet joint-detection model.
10. device according to claim 8 is it is characterised in that described number of people safety helmet joint-detection model includes the number of people
Detection model and safety helmet position detection model, described number of people safety helmet united state detection module includes:
Number of people region detection unit, obtains people for carrying out detection according to described number of people detection model to described live video image
Head region；
Safety helmet region detection unit, carries out detecting to described live video image according to described safety helmet position detection model
To safety helmet region；
Number of people safety helmet united state determining unit, for the distance according to described people's head region and safety helmet region and faying surface
Amass and obtain number of people safety helmet united state.
11. devices according to claim 10 are it is characterised in that described device also includes:
Position model sets up module, and for obtaining training video image, described training video image is included containing examined object
Positive sample image, do not contain the negative sample image of examined object, described examined object is human body, in the number of people, safety helmet
At least one, obtain the markup information set of described training video image, described markup information set includes human region mark
Note information, number of people area marking information and safety helmet area marking information, obtain current training mould from described markup information set
The corresponding current markup information of type, current markup information forms current training data with training video image, by described current instruction
Practice the current training pattern of data input, described current training data is trained obtain based on the method for training study corresponding
Current detection model, described current training pattern includes safety helmet position detection training pattern, position of human body detection training pattern
With the number of people detect training pattern, described current detection model include described safety helmet detection model, position of human body detection model and
Number of people detection model.
12. devices according to claim 8 are it is characterised in that described device also includes:
Safety helmet state detection module, for the safety helmet state-detection model using training study foundation to described live video
Image carries out detection and obtains safety helmet current state, judges whether safety helmet damages according to described safety helmet current state, if
Safety helmet damage, then safety helmet wear not verified.
13. devices according to claim 8 are it is characterised in that described acquisition module is additionally operable to obtain described live video
The corresponding camera identification of image, described device also includes:
The direct judge module in region worn by safety helmet, for judging whether the corresponding shooting area of described camera identification is panorama
Region worn by safety helmet, if it is, judging to wear region in safety helmet, is directly entered number of people safety helmet united state detection mould
Block, otherwise enters safety helmet and wears region detection module；
Safety helmet is worn region detection module and is additionally operable to wear area according to the described camera identification corresponding current safety cap of acquisition
Domain, judges whether described position of human body wears region in described current safety cap；
Described safety helmet type detection module is additionally operable to obtain corresponding current region safety helmet mark according to described camera identification
Quasi- type, if described safety helmet type is mated with described current region safety helmet type, described safety helmet type symbol
Close regulation, otherwise described safety helmet type mismatch closes regulation.
Device described in 14. according to Claim 8 or 12 is it is characterised in that described device also includes:
Prompting module, wears not verified reminder message to terminal for sending safety helmet, described reminder message includes regarding
Frequently, image, at least one in voice, information the reason described reminder message cap safe to carry wears not verified.
A kind of 15. photographic head are it is characterised in that described photographic head includes the safety described in any one of the claims 8 to 14
The detection means that cap is worn.
A kind of 16. servers are it is characterised in that described server includes the safety described in any one of the claims 8 to 14
The detection means that cap is worn.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610778655.0A CN106372662B (en) | 2016-08-30 | 2016-08-30 | Detection method and device for wearing of safety helmet, camera and server |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610778655.0A CN106372662B (en) | 2016-08-30 | 2016-08-30 | Detection method and device for wearing of safety helmet, camera and server |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106372662A true CN106372662A (en) | 2017-02-01 |
| CN106372662B CN106372662B (en) | 2020-04-28 |
Family
ID=57902358
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610778655.0A Active CN106372662B (en) | 2016-08-30 | 2016-08-30 | Detection method and device for wearing of safety helmet, camera and server |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106372662B (en) |
Cited By (61)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107563281A (en) * | 2017-07-24 | 2018-01-09 | 南京邮电大学 | A kind of construction site personal security hidden danger monitoring method based on deep learning |
| CN107679524A (en) * | 2017-10-31 | 2018-02-09 | 天津天地伟业信息系统集成有限公司 | A kind of detection method of the safety cap wear condition based on video |
| CN108052900A (en) * | 2017-12-12 | 2018-05-18 | 成都睿码科技有限责任公司 | A kind of method by monitor video automatic decision dressing specification |
| CN108197576A (en) * | 2018-01-08 | 2018-06-22 | 安徽炬视科技有限公司 | Safety cap based on DSP technologies wears detection device and its detection method |
| CN108304831A (en) * | 2018-03-15 | 2018-07-20 | 广东工业大学 | A kind of method and device that monitoring worker safety helmet is worn |
| CN108319934A (en) * | 2018-03-20 | 2018-07-24 | 武汉倍特威视系统有限公司 | Safety cap wear condition detection method based on video stream data |
| CN108319926A (en) * | 2018-02-12 | 2018-07-24 | 安徽金禾软件股份有限公司 | A kind of the safety cap wearing detecting system and detection method of building-site |
| CN108388873A (en) * | 2018-03-01 | 2018-08-10 | 路志宏 | A kind of supervisory systems of machine vision, method and client computer, storage medium |
| CN108399389A (en) * | 2018-03-01 | 2018-08-14 | 路志宏 | Multimachine supervisory systems, method and the client computer of machine vision, server, storage medium |
| CN108460358A (en) * | 2018-03-20 | 2018-08-28 | 武汉倍特威视系统有限公司 | Safety cap recognition methods based on video stream data |
| CN108537256A (en) * | 2018-03-26 | 2018-09-14 | 北京智芯原动科技有限公司 | A kind of safety cap wears recognition methods and device |
| CN108596111A (en) * | 2018-04-26 | 2018-09-28 | 四川隧唐科技股份有限公司 | Safety cap recognition methods and device |
| CN108629767A (en) * | 2018-04-28 | 2018-10-09 | Oppo广东移动通信有限公司 | A kind of method, device and mobile terminal of scene detection |
| CN108921159A (en) * | 2018-07-26 | 2018-11-30 | 北京百度网讯科技有限公司 | Method and apparatus for detecting the wear condition of safety cap |
| CN109255298A (en) * | 2018-08-07 | 2019-01-22 | 南京工业大学 | Safety helmet detection method and system in dynamic background |
| CN109376676A (en) * | 2018-11-01 | 2019-02-22 | 哈尔滨工业大学 | Highway engineering site operation personnel safety method for early warning based on unmanned aerial vehicle platform |
| CN109447176A (en) * | 2018-11-09 | 2019-03-08 | 广东工业大学 | Bursting work personnel detection method, device, system, medium and server |
| CN109557935A (en) * | 2018-11-01 | 2019-04-02 | 广州圣安环保科技有限公司 | A kind of safety monitoring during construction method and system based on unmanned plane |
| CN109697430A (en) * | 2018-12-28 | 2019-04-30 | 成都思晗科技股份有限公司 | The detection method that working region safety cap based on image recognition is worn |
| CN109711274A (en) * | 2018-12-05 | 2019-05-03 | 斑马网络技术有限公司 | Vehicle checking method, device, equipment and storage medium |
| CN109753898A (en) * | 2018-12-21 | 2019-05-14 | 中国三峡建设管理有限公司 | A kind of safety cap recognition methods and device |
| US10296794B2 (en) | 2016-12-20 | 2019-05-21 | Jayant Rtti | On-demand artificial intelligence and roadway stewardship system |
| CN109886205A (en) * | 2019-02-25 | 2019-06-14 | 苏州清研微视电子科技有限公司 | Safety belt method of real-time and system |
| CN110020598A (en) * | 2019-02-28 | 2019-07-16 | 中电海康集团有限公司 | A kind of method and device based on foreign matter on deep learning detection electric pole |
| CN110060299A (en) * | 2019-04-18 | 2019-07-26 | 中国测绘科学研究院 | Danger source identifies and positions method in passway for transmitting electricity based on binocular vision technology |
| CN110188724A (en) * | 2019-06-05 | 2019-08-30 | 中冶赛迪重庆信息技术有限公司 | The method and system of safety cap positioning and color identification based on deep learning |
| CN110210274A (en) * | 2018-02-28 | 2019-09-06 | 杭州海康威视数字技术股份有限公司 | Safety cap detection method, device and computer readable storage medium |
| WO2019175686A1 (en) | 2018-03-12 | 2019-09-19 | Ratti Jayant | On-demand artificial intelligence and roadway stewardship system |
| CN110426746A (en) * | 2019-06-28 | 2019-11-08 | 北京市政建设集团有限责任公司 | Head-mount formula safety equipment wears detection method |
| CN110458075A (en) * | 2019-08-05 | 2019-11-15 | 北京泰豪信息科技有限公司 | Detection method, storage medium, detection device and the detection system that safety cap is worn |
| CN110618436A (en) * | 2019-04-04 | 2019-12-27 | 中国石油大学(北京) | Inspection method, device and equipment based on instant positioning and map construction |
| CN110781706A (en) * | 2018-07-30 | 2020-02-11 | 杭州海康威视系统技术有限公司 | Safety belt wearing detection method and device and computer readable storage medium |
| CN111047280A (en) * | 2019-11-22 | 2020-04-21 | 三一重工股份有限公司 | Method and device for detecting objects in plant and computer readable storage medium |
| CN111083441A (en) * | 2019-12-18 | 2020-04-28 | 广州穗能通能源科技有限责任公司 | Construction site monitoring method and device, computer equipment and storage medium |
| CN111079722A (en) * | 2020-03-23 | 2020-04-28 | 杭州鲁尔物联科技有限公司 | Hoisting process personnel safety monitoring method and system |
| CN111191586A (en) * | 2019-12-30 | 2020-05-22 | 安徽小眯当家信息技术有限公司 | Method and system for inspecting wearing condition of safety helmet of personnel in construction site |
| CN111191581A (en) * | 2019-12-27 | 2020-05-22 | 深圳供电局有限公司 | Safety helmet detection method and device based on electric power construction and computer equipment |
| CN111199200A (en) * | 2019-12-27 | 2020-05-26 | 深圳供电局有限公司 | Wearing detection method and device based on electric protection equipment and computer equipment |
| CN111222420A (en) * | 2019-12-24 | 2020-06-02 | 重庆市通信产业服务有限公司 | FTP protocol-based low-bandwidth-requirement helmet identification method |
| CN111259763A (en) * | 2020-01-13 | 2020-06-09 | 华雁智能科技(集团)股份有限公司 | Target detection method and device, electronic equipment and readable storage medium |
| CN111259855A (en) * | 2020-02-09 | 2020-06-09 | 天津博宜特科技有限公司 | Mobile safety helmet wearing detection method based on deep learning |
| CN111401310A (en) * | 2020-04-08 | 2020-07-10 | 天津中科智能识别产业技术研究院有限公司 | Kitchen health safety supervision and management method based on artificial intelligence |
| CN111414873A (en) * | 2020-03-26 | 2020-07-14 | 广州粤建三和软件股份有限公司 | Alarm prompting method, device and alarm system based on wearing state of safety helmet |
| CN111428641A (en) * | 2020-03-24 | 2020-07-17 | 深圳供电局有限公司 | Secure dressing detection method and device, computer equipment and readable storage medium |
| CN111444869A (en) * | 2020-03-31 | 2020-07-24 | 高新兴科技集团股份有限公司 | Method and device for identifying wearing state of mask and computer equipment |
| CN111539276A (en) * | 2020-04-14 | 2020-08-14 | 国家电网有限公司 | Method for detecting safety helmet in real time in power scene |
| CN111666920A (en) * | 2020-06-24 | 2020-09-15 | 浙江大华技术股份有限公司 | Target object wearing detection method and device, storage medium and electronic device |
| CN111738222A (en) * | 2020-07-28 | 2020-10-02 | 合肥小步智能科技有限公司 | Safety helmet detection device and method |
| CN111860181A (en) * | 2020-06-23 | 2020-10-30 | 国网山东省电力公司淄博供电公司 | Distributed safety helmet identification system |
| CN111931653A (en) * | 2020-08-11 | 2020-11-13 | 沈阳帝信人工智能产业研究院有限公司 | Safety monitoring method and device, electronic equipment and readable storage medium |
| CN112101288A (en) * | 2020-09-25 | 2020-12-18 | 北京百度网讯科技有限公司 | Method, device and equipment for detecting wearing of safety helmet and storage medium |
| US10959479B1 (en) | 2019-12-27 | 2021-03-30 | Robert Bosch Gmbh | Apparatus and warning system for intelligent helmet |
| CN113033481A (en) * | 2021-04-20 | 2021-06-25 | 湖北工业大学 | Method for detecting hand-held stick combined with aspect ratio-first order fully-convolved object detection (FCOS) algorithm |
| CN113139426A (en) * | 2021-03-12 | 2021-07-20 | 浙江智慧视频安防创新中心有限公司 | Detection method and device for wearing safety helmet, storage medium and terminal |
| CN113158851A (en) * | 2021-04-07 | 2021-07-23 | 浙江大华技术股份有限公司 | Wearing safety helmet detection method and device and computer storage medium |
| CN113177461A (en) * | 2021-04-25 | 2021-07-27 | 台州智必安科技有限责任公司 | Passenger attachment detection method and system based on hardware front end |
| CN113283296A (en) * | 2021-04-20 | 2021-08-20 | 晋城鸿智纳米光机电研究院有限公司 | Helmet wearing detection method, electronic device and storage medium |
| CN113455771A (en) * | 2021-08-20 | 2021-10-01 | 葛洲坝易普力四川爆破工程有限公司 | Intelligent safety helmet based on 5G communication for blasting field and management method thereof |
| CN116740753A (en) * | 2023-04-20 | 2023-09-12 | 安徽大学 | Target detection and tracking method and system based on improved YOLOv5 and deep SORT |
| CN116824618A (en) * | 2022-11-30 | 2023-09-29 | 慧之安信息技术股份有限公司 | Anti-cutting glove wearing detection method based on deep learning |
| CN118379688A (en) * | 2024-06-26 | 2024-07-23 | 沃太能源股份有限公司 | Helmet wearing identification method and device, storage medium and electronic terminal |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101521805A (en) * | 2009-02-27 | 2009-09-02 | 镇江石鼓文智能化系统开发有限公司 | Helmet-site detecting system on construction field and detecting method thereof |
| JP2010211427A (en) * | 2009-03-10 | 2010-09-24 | Omron Corp | Safe driving monitor |
| CN102136076A (en) * | 2011-03-14 | 2011-07-27 | 徐州中矿大华洋通信设备有限公司 | Method for positioning and tracing underground personnel of coal mine based on safety helmet detection |
| CN203012909U (en) * | 2012-11-16 | 2013-06-19 | 上海工业自动化仪表研究院 | Worker safety cap identifying system for industrial site |
| CN104036575A (en) * | 2014-07-01 | 2014-09-10 | 江苏省南京市公路管理处公路科学研究所 | Safety helmet wearing condition monitoring method on construction site |
| CN104063722A (en) * | 2014-07-15 | 2014-09-24 | 国家电网公司 | Safety helmet identification method integrating HOG human body target detection and SVM classifier |
| CN104318732A (en) * | 2014-10-27 | 2015-01-28 | 国网冀北电力有限公司张家口供电公司 | Transformer substation field worker monitoring and management system and method based on video analysis and RFID |
| CN104504369A (en) * | 2014-12-12 | 2015-04-08 | 无锡北邮感知技术产业研究院有限公司 | Wearing condition detection method for safety helmets |
| WO2015076152A1 (en) * | 2013-11-20 | 2015-05-28 | 日本電気株式会社 | Helmet use assessment method, helmet use assessment system, helmet use assessment apparatus and program |
| CN104902227A (en) * | 2015-05-06 | 2015-09-09 | 南京第五十五所技术开发有限公司 | Substation helmet wearing condition video monitoring system |
| CN105469038A (en) * | 2015-11-19 | 2016-04-06 | 国网福建省电力有限公司泉州供电公司 | Safety cap video detection method for electric power switching station |
| CN105469069A (en) * | 2015-12-09 | 2016-04-06 | 上海工业自动化仪表研究院 | Safety helmet video detection method for production line data acquisition terminal |
| CN105760847A (en) * | 2016-03-01 | 2016-07-13 | 江苏大学 | Visual detection method for detecting whether motor cyclist wears helmet or not |
-
2016
- 2016-08-30 CN CN201610778655.0A patent/CN106372662B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101521805A (en) * | 2009-02-27 | 2009-09-02 | 镇江石鼓文智能化系统开发有限公司 | Helmet-site detecting system on construction field and detecting method thereof |
| JP2010211427A (en) * | 2009-03-10 | 2010-09-24 | Omron Corp | Safe driving monitor |
| CN102136076A (en) * | 2011-03-14 | 2011-07-27 | 徐州中矿大华洋通信设备有限公司 | Method for positioning and tracing underground personnel of coal mine based on safety helmet detection |
| CN203012909U (en) * | 2012-11-16 | 2013-06-19 | 上海工业自动化仪表研究院 | Worker safety cap identifying system for industrial site |
| WO2015076152A1 (en) * | 2013-11-20 | 2015-05-28 | 日本電気株式会社 | Helmet use assessment method, helmet use assessment system, helmet use assessment apparatus and program |
| CN104036575A (en) * | 2014-07-01 | 2014-09-10 | 江苏省南京市公路管理处公路科学研究所 | Safety helmet wearing condition monitoring method on construction site |
| CN104063722A (en) * | 2014-07-15 | 2014-09-24 | 国家电网公司 | Safety helmet identification method integrating HOG human body target detection and SVM classifier |
| CN104318732A (en) * | 2014-10-27 | 2015-01-28 | 国网冀北电力有限公司张家口供电公司 | Transformer substation field worker monitoring and management system and method based on video analysis and RFID |
| CN104504369A (en) * | 2014-12-12 | 2015-04-08 | 无锡北邮感知技术产业研究院有限公司 | Wearing condition detection method for safety helmets |
| CN104902227A (en) * | 2015-05-06 | 2015-09-09 | 南京第五十五所技术开发有限公司 | Substation helmet wearing condition video monitoring system |
| CN105469038A (en) * | 2015-11-19 | 2016-04-06 | 国网福建省电力有限公司泉州供电公司 | Safety cap video detection method for electric power switching station |
| CN105469069A (en) * | 2015-12-09 | 2016-04-06 | 上海工业自动化仪表研究院 | Safety helmet video detection method for production line data acquisition terminal |
| CN105760847A (en) * | 2016-03-01 | 2016-07-13 | 江苏大学 | Visual detection method for detecting whether motor cyclist wears helmet or not |
Non-Patent Citations (8)
| Title |
|---|
| RATTAPOOM WARANUSAST等: ""Machine vision techniques for motorcycle safety helmet detection"", 《2013 28TH INTERNATIONAL CONFERENCE ON IMAGE AND VISION COMPUTING NEW ZEALAND (IVCNZ 2013)》 * |
| ROMUERE SILVA等: ""Helmet Detection on Motorcyclists Using Image Descriptors and Classifiers"", 《2014 27TH SIBGRAPI CONFERENCE ON GRAPHICS,PATTERNS AND IMAGES》 * |
| SHREYA BHAGAT等: ""Cascade Classifier based Helmet Detection using OpenCV in Image Processing"", 《HTTPS://WWW.RESEARCHGATE.NET/PUBLICATION/327642532》 * |
| 冯国臣等: ""基于机器视觉的安全帽自动识别技术研究"", 《机械设计与制造工程》 * |
| 冯杰: ""图像识别技术在换流站监控系统中的应用研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 刘云波等: ""施工现场安全帽佩戴情况监控技术研究"", 《电子科技》 * |
| 李太华等: ""基于安全帽颜色识别的人员身份认证算法在变电站的应用"", 《电脑知识与技术》 * |
| 李潇: ""视频识别在HSE监控平台中的研究与应用"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (80)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10296794B2 (en) | 2016-12-20 | 2019-05-21 | Jayant Rtti | On-demand artificial intelligence and roadway stewardship system |
| CN107563281A (en) * | 2017-07-24 | 2018-01-09 | 南京邮电大学 | A kind of construction site personal security hidden danger monitoring method based on deep learning |
| CN107679524A (en) * | 2017-10-31 | 2018-02-09 | 天津天地伟业信息系统集成有限公司 | A kind of detection method of the safety cap wear condition based on video |
| CN108052900A (en) * | 2017-12-12 | 2018-05-18 | 成都睿码科技有限责任公司 | A kind of method by monitor video automatic decision dressing specification |
| CN108197576A (en) * | 2018-01-08 | 2018-06-22 | 安徽炬视科技有限公司 | Safety cap based on DSP technologies wears detection device and its detection method |
| CN108197576B (en) * | 2018-01-08 | 2020-04-17 | 安徽炬视科技有限公司 | Safety helmet wearing detection device based on DSP technology and detection method thereof |
| CN108319926A (en) * | 2018-02-12 | 2018-07-24 | 安徽金禾软件股份有限公司 | A kind of the safety cap wearing detecting system and detection method of building-site |
| CN110210274A (en) * | 2018-02-28 | 2019-09-06 | 杭州海康威视数字技术股份有限公司 | Safety cap detection method, device and computer readable storage medium |
| CN108388873A (en) * | 2018-03-01 | 2018-08-10 | 路志宏 | A kind of supervisory systems of machine vision, method and client computer, storage medium |
| CN108399389A (en) * | 2018-03-01 | 2018-08-14 | 路志宏 | Multimachine supervisory systems, method and the client computer of machine vision, server, storage medium |
| CN108399389B (en) * | 2018-03-01 | 2020-05-08 | 路志宏 | Multi-machine monitoring system and method for machine vision, client, server and storage medium |
| WO2019175686A1 (en) | 2018-03-12 | 2019-09-19 | Ratti Jayant | On-demand artificial intelligence and roadway stewardship system |
| CN108304831A (en) * | 2018-03-15 | 2018-07-20 | 广东工业大学 | A kind of method and device that monitoring worker safety helmet is worn |
| CN108304831B (en) * | 2018-03-15 | 2022-03-22 | 广东工业大学 | Method and device for monitoring wearing of safety helmet of worker |
| CN108319934A (en) * | 2018-03-20 | 2018-07-24 | 武汉倍特威视系统有限公司 | Safety cap wear condition detection method based on video stream data |
| CN108460358A (en) * | 2018-03-20 | 2018-08-28 | 武汉倍特威视系统有限公司 | Safety cap recognition methods based on video stream data |
| CN108537256A (en) * | 2018-03-26 | 2018-09-14 | 北京智芯原动科技有限公司 | A kind of safety cap wears recognition methods and device |
| CN108537256B (en) * | 2018-03-26 | 2021-09-17 | 北京智芯原动科技有限公司 | Method and device for identifying wearing of safety helmet |
| CN108596111A (en) * | 2018-04-26 | 2018-09-28 | 四川隧唐科技股份有限公司 | Safety cap recognition methods and device |
| CN108629767A (en) * | 2018-04-28 | 2018-10-09 | Oppo广东移动通信有限公司 | A kind of method, device and mobile terminal of scene detection |
| CN108921159A (en) * | 2018-07-26 | 2018-11-30 | 北京百度网讯科技有限公司 | Method and apparatus for detecting the wear condition of safety cap |
| CN108921159B (en) * | 2018-07-26 | 2022-09-27 | 北京百度网讯科技有限公司 | Method and device for detecting wearing condition of safety helmet |
| CN110781706A (en) * | 2018-07-30 | 2020-02-11 | 杭州海康威视系统技术有限公司 | Safety belt wearing detection method and device and computer readable storage medium |
| CN109255298A (en) * | 2018-08-07 | 2019-01-22 | 南京工业大学 | Safety helmet detection method and system in dynamic background |
| CN109557935A (en) * | 2018-11-01 | 2019-04-02 | 广州圣安环保科技有限公司 | A kind of safety monitoring during construction method and system based on unmanned plane |
| CN109376676A (en) * | 2018-11-01 | 2019-02-22 | 哈尔滨工业大学 | Highway engineering site operation personnel safety method for early warning based on unmanned aerial vehicle platform |
| CN109447176A (en) * | 2018-11-09 | 2019-03-08 | 广东工业大学 | Bursting work personnel detection method, device, system, medium and server |
| CN109711274A (en) * | 2018-12-05 | 2019-05-03 | 斑马网络技术有限公司 | Vehicle checking method, device, equipment and storage medium |
| CN109753898A (en) * | 2018-12-21 | 2019-05-14 | 中国三峡建设管理有限公司 | A kind of safety cap recognition methods and device |
| CN109697430A (en) * | 2018-12-28 | 2019-04-30 | 成都思晗科技股份有限公司 | The detection method that working region safety cap based on image recognition is worn |
| CN109886205A (en) * | 2019-02-25 | 2019-06-14 | 苏州清研微视电子科技有限公司 | Safety belt method of real-time and system |
| CN109886205B (en) * | 2019-02-25 | 2023-08-08 | 苏州清研微视电子科技有限公司 | Real-time safety belt monitoring method and system |
| CN110020598B (en) * | 2019-02-28 | 2022-04-15 | 中电海康集团有限公司 | Method and device for detecting foreign matters on telegraph pole based on deep learning |
| CN110020598A (en) * | 2019-02-28 | 2019-07-16 | 中电海康集团有限公司 | A kind of method and device based on foreign matter on deep learning detection electric pole |
| CN110618436A (en) * | 2019-04-04 | 2019-12-27 | 中国石油大学(北京) | Inspection method, device and equipment based on instant positioning and map construction |
| CN110060299A (en) * | 2019-04-18 | 2019-07-26 | 中国测绘科学研究院 | Danger source identifies and positions method in passway for transmitting electricity based on binocular vision technology |
| CN110188724B (en) * | 2019-06-05 | 2023-02-28 | 中冶赛迪信息技术(重庆)有限公司 | Method and system for helmet positioning and color recognition based on deep learning |
| CN110188724A (en) * | 2019-06-05 | 2019-08-30 | 中冶赛迪重庆信息技术有限公司 | The method and system of safety cap positioning and color identification based on deep learning |
| CN110426746A (en) * | 2019-06-28 | 2019-11-08 | 北京市政建设集团有限责任公司 | Head-mount formula safety equipment wears detection method |
| CN110458075A (en) * | 2019-08-05 | 2019-11-15 | 北京泰豪信息科技有限公司 | Detection method, storage medium, detection device and the detection system that safety cap is worn |
| CN110458075B (en) * | 2019-08-05 | 2023-08-25 | 北京泰豪信息科技有限公司 | Method, storage medium, device and system for detecting wearing of safety helmet |
| CN111047280A (en) * | 2019-11-22 | 2020-04-21 | 三一重工股份有限公司 | Method and device for detecting objects in plant and computer readable storage medium |
| CN111083441A (en) * | 2019-12-18 | 2020-04-28 | 广州穗能通能源科技有限责任公司 | Construction site monitoring method and device, computer equipment and storage medium |
| CN111222420A (en) * | 2019-12-24 | 2020-06-02 | 重庆市通信产业服务有限公司 | FTP protocol-based low-bandwidth-requirement helmet identification method |
| CN111199200A (en) * | 2019-12-27 | 2020-05-26 | 深圳供电局有限公司 | Wearing detection method and device based on electric protection equipment and computer equipment |
| CN111191581B (en) * | 2019-12-27 | 2024-04-12 | 深圳供电局有限公司 | Safety helmet detection method and device based on electric power construction and computer equipment |
| CN111191581A (en) * | 2019-12-27 | 2020-05-22 | 深圳供电局有限公司 | Safety helmet detection method and device based on electric power construction and computer equipment |
| US10959479B1 (en) | 2019-12-27 | 2021-03-30 | Robert Bosch Gmbh | Apparatus and warning system for intelligent helmet |
| CN111191586A (en) * | 2019-12-30 | 2020-05-22 | 安徽小眯当家信息技术有限公司 | Method and system for inspecting wearing condition of safety helmet of personnel in construction site |
| CN111191586B (en) * | 2019-12-30 | 2020-11-27 | 安徽小眯当家信息技术有限公司 | Method and system for inspecting wearing condition of safety helmet of personnel in construction site |
| CN111259763B (en) * | 2020-01-13 | 2024-02-02 | 华雁智能科技(集团)股份有限公司 | Target detection method, target detection device, electronic equipment and readable storage medium |
| CN111259763A (en) * | 2020-01-13 | 2020-06-09 | 华雁智能科技(集团)股份有限公司 | Target detection method and device, electronic equipment and readable storage medium |
| CN111259855A (en) * | 2020-02-09 | 2020-06-09 | 天津博宜特科技有限公司 | Mobile safety helmet wearing detection method based on deep learning |
| CN111079722A (en) * | 2020-03-23 | 2020-04-28 | 杭州鲁尔物联科技有限公司 | Hoisting process personnel safety monitoring method and system |
| CN111428641A (en) * | 2020-03-24 | 2020-07-17 | 深圳供电局有限公司 | Secure dressing detection method and device, computer equipment and readable storage medium |
| CN111414873A (en) * | 2020-03-26 | 2020-07-14 | 广州粤建三和软件股份有限公司 | Alarm prompting method, device and alarm system based on wearing state of safety helmet |
| CN111414873B (en) * | 2020-03-26 | 2021-04-30 | 广州粤建三和软件股份有限公司 | Alarm prompting method, device and alarm system based on wearing state of safety helmet |
| CN111444869A (en) * | 2020-03-31 | 2020-07-24 | 高新兴科技集团股份有限公司 | Method and device for identifying wearing state of mask and computer equipment |
| CN111401310B (en) * | 2020-04-08 | 2023-08-29 | 天津中科智能识别产业技术研究院有限公司 | Kitchen sanitation safety supervision and management method based on artificial intelligence |
| CN111401310A (en) * | 2020-04-08 | 2020-07-10 | 天津中科智能识别产业技术研究院有限公司 | Kitchen health safety supervision and management method based on artificial intelligence |
| CN111539276A (en) * | 2020-04-14 | 2020-08-14 | 国家电网有限公司 | Method for detecting safety helmet in real time in power scene |
| CN111860181A (en) * | 2020-06-23 | 2020-10-30 | 国网山东省电力公司淄博供电公司 | Distributed safety helmet identification system |
| CN111666920B (en) * | 2020-06-24 | 2023-09-01 | 浙江大华技术股份有限公司 | Target article wearing detection method and device, storage medium and electronic device |
| CN111666920A (en) * | 2020-06-24 | 2020-09-15 | 浙江大华技术股份有限公司 | Target object wearing detection method and device, storage medium and electronic device |
| CN111738222A (en) * | 2020-07-28 | 2020-10-02 | 合肥小步智能科技有限公司 | Safety helmet detection device and method |
| CN111931653B (en) * | 2020-08-11 | 2024-06-11 | 沈阳帝信人工智能产业研究院有限公司 | Security monitoring method, device, electronic equipment and readable storage medium |
| CN111931653A (en) * | 2020-08-11 | 2020-11-13 | 沈阳帝信人工智能产业研究院有限公司 | Safety monitoring method and device, electronic equipment and readable storage medium |
| CN112101288B (en) * | 2020-09-25 | 2024-02-13 | 北京百度网讯科技有限公司 | Method, device, equipment and storage medium for detecting wearing of safety helmet |
| CN112101288A (en) * | 2020-09-25 | 2020-12-18 | 北京百度网讯科技有限公司 | Method, device and equipment for detecting wearing of safety helmet and storage medium |
| CN113139426A (en) * | 2021-03-12 | 2021-07-20 | 浙江智慧视频安防创新中心有限公司 | Detection method and device for wearing safety helmet, storage medium and terminal |
| WO2022213552A1 (en) * | 2021-04-07 | 2022-10-13 | Zhejiang Dahua Technology Co., Ltd. | Systems and methods for helmet detection |
| CN113158851B (en) * | 2021-04-07 | 2022-08-09 | 浙江大华技术股份有限公司 | Wearing safety helmet detection method and device and computer storage medium |
| CN113158851A (en) * | 2021-04-07 | 2021-07-23 | 浙江大华技术股份有限公司 | Wearing safety helmet detection method and device and computer storage medium |
| CN113033481A (en) * | 2021-04-20 | 2021-06-25 | 湖北工业大学 | Method for detecting hand-held stick combined with aspect ratio-first order fully-convolved object detection (FCOS) algorithm |
| CN113283296A (en) * | 2021-04-20 | 2021-08-20 | 晋城鸿智纳米光机电研究院有限公司 | Helmet wearing detection method, electronic device and storage medium |
| CN113177461A (en) * | 2021-04-25 | 2021-07-27 | 台州智必安科技有限责任公司 | Passenger attachment detection method and system based on hardware front end |
| CN113455771A (en) * | 2021-08-20 | 2021-10-01 | 葛洲坝易普力四川爆破工程有限公司 | Intelligent safety helmet based on 5G communication for blasting field and management method thereof |
| CN116824618A (en) * | 2022-11-30 | 2023-09-29 | 慧之安信息技术股份有限公司 | Anti-cutting glove wearing detection method based on deep learning |
| CN116740753A (en) * | 2023-04-20 | 2023-09-12 | 安徽大学 | Target detection and tracking method and system based on improved YOLOv5 and deep SORT |
| CN118379688A (en) * | 2024-06-26 | 2024-07-23 | 沃太能源股份有限公司 | Helmet wearing identification method and device, storage medium and electronic terminal |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106372662B (en) | 2020-04-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106372662A (en) | Helmet wearing detection method and device, camera, and server | |
| CN110502965B (en) | Construction safety helmet wearing monitoring method based on computer vision human body posture estimation | |
| CN108319926A (en) | A kind of the safety cap wearing detecting system and detection method of building-site | |
| CN108009473A (en) | Based on goal behavior attribute video structural processing method, system and storage device | |
| CN108053427A (en) | A kind of modified multi-object tracking method, system and device based on KCF and Kalman | |
| CN106295567A (en) | The localization method of a kind of key point and terminal | |
| CN109253888A (en) | Detection method and system for vehicle vehicle condition | |
| CN111126153B (en) | Safety monitoring method, system, server and storage medium based on deep learning | |
| CN104166841A (en) | Rapid detection identification method for specified pedestrian or vehicle in video monitoring network | |
| CN107643125A (en) | The determination method and apparatus of equipment fault | |
| CN113516076A (en) | Improved lightweight YOLO v4 safety protection detection method based on attention mechanism | |
| CN110516518A (en) | A kind of illegal manned detection method of non-motor vehicle, device and electronic equipment | |
| CN107609544A (en) | A kind of detection method and device | |
| CN113361354A (en) | Track component inspection method and device, computer equipment and storage medium | |
| CN112183472A (en) | Method for detecting whether test field personnel wear work clothes or not based on improved RetinaNet | |
| CN109670517A (en) | Object detection method, device, electronic equipment and target detection model | |
| CN108960124A (en) | The image processing method and device identified again for pedestrian | |
| CN111753587A (en) | Method and device for detecting falling to ground | |
| CN116863297A (en) | Monitoring method, device, system, equipment and medium based on electronic fence | |
| CN109800656B (en) | Positioning method and related product | |
| CN108764264A (en) | Smog detection method, smoke detection system and computer installation | |
| CN113516082A (en) | Detection method and device of safety helmet, computer equipment and storage medium | |
| CN117853979A (en) | Method and device for detecting wearing of safety helmet, electronic equipment and storage medium | |
| CN112001336A (en) | Pedestrian boundary crossing alarm method, device, equipment and system | |
| CN113361968B (en) | Power grid infrastructure worker safety risk assessment method based on artificial intelligence and big data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |