CN115311820A - Intelligent security system near water - Google Patents
Intelligent security system near water Download PDFInfo
- Publication number
- CN115311820A CN115311820A CN202210811561.4A CN202210811561A CN115311820A CN 115311820 A CN115311820 A CN 115311820A CN 202210811561 A CN202210811561 A CN 202210811561A CN 115311820 A CN115311820 A CN 115311820A
- Authority
- CN
- China
- Prior art keywords
- water
- personnel
- security system
- video acquisition
- intelligent security
- 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.)
- Pending
Links
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 57
- 238000001514 detection method Methods 0.000 claims abstract description 45
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 29
- 206010013647 Drowning Diseases 0.000 claims description 24
- 230000007246 mechanism Effects 0.000 claims description 9
- 238000000034 method Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000003213 activating effect Effects 0.000 claims 1
- 230000001681 protective effect Effects 0.000 description 7
- 230000000903 blocking effect Effects 0.000 description 6
- 238000005457 optimization Methods 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 4
- ATJFFYVFTNAWJD-UHFFFAOYSA-N Tin Chemical compound [Sn] ATJFFYVFTNAWJD-UHFFFAOYSA-N 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 239000011521 glass Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/08—Alarms for ensuring the safety of persons responsive to the presence of persons in a body of water, e.g. a swimming pool; responsive to an abnormal condition of a body of water
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
- G08B13/02—Mechanical actuation
- G08B13/12—Mechanical actuation by the breaking or disturbance of stretched cords or wires
- G08B13/122—Mechanical actuation by the breaking or disturbance of stretched cords or wires for a perimeter fence
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/08—Alarms for ensuring the safety of persons responsive to the presence of persons in a body of water, e.g. a swimming pool; responsive to an abnormal condition of a body of water
- G08B21/088—Alarms for ensuring the safety of persons responsive to the presence of persons in a body of water, e.g. a swimming pool; responsive to an abnormal condition of a body of water by monitoring a device worn by the person, e.g. a bracelet attached to the swimmer
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Alarm Systems (AREA)
Abstract
The invention relates to the technical field of intelligent security, in particular to an intelligent security system for near water, which comprises a detection module, an identification module falling into water and a classification module, wherein the detection module, the identification module falling into water and the classification module all need to operate on computer equipment, the detection module and the identification module falling into water all need to be connected with a plurality of video acquisition devices, the video acquisition devices are integrated with alarms and auxiliary positioning devices, the video acquisition devices are provided with power connectors through power lines, the video acquisition devices are positioned on one side of a near water area, and the detection module and the identification module falling into water utilize a YOLOv5 target detection algorithm to identify and detect personnel, perform Deepsort tracking on the personnel, calculate the number of the personnel by using a counter and record the personnel in the system; the invention can solve the problems of low accuracy and incomplete function of the existing system, is easy to generate a large number of false alarms and the like, and effectively improves the accuracy.
Description
Technical Field
The invention relates to a security system, in particular to an intelligent near-water security system, and belongs to the technical field of intelligent security.
Background
In shore-based sites such as wharfs and the like, correct wearing of a life jacket by workers is an important measure for ensuring the safety of the workers, real-time alarming of falling water of the workers is the only measure for saving the lives of the workers, in the production process, a lot of uncertainty exists, a near-water intelligent security system can monitor and guarantee the life safety of the workers, and therefore the life jacket safety monitoring system is required to be considered from the whole world with all-weather high precision, the falling water identification function is combined, the problems of low accuracy, incomplete functions and the like in the prior art are solved, and the life safety of the workers can be greatly improved and guaranteed.
As a Chinese application patent document, an application document with publication number CN111028480A, discloses a drowning detection and alarm system, which can collect and monitor video frame data of a camera at the front end of a water area in real time through a video collection module, a person drowning detection module defines a detection area for the video frame data, then a person drowning detection module is used for carrying out person detection on the video frame data in the defined detection area, once the drowning condition of a person is found, an alarm signal is sent in time, and the drowning information of the person is sent to a background server; the function of the method is single relatively, only a single alarm can be realized, the functionality is poor, the detection method is general, no pertinence exists, and no algorithm part is described.
For example, a Chinese application patent document, publication No. CN111210464A, discloses a personnel drowning alarm system and method based on convolutional neural network and image fusion, which performs target detection through a convolutional neural network module to judge whether a dangerous condition of personnel drowning occurs; the control alarm module feeds back the judgment result of the convolutional neural network module to the control alarm module, and when a dangerous case occurs, the alarm is started and the information of falling water is stored; the image display module displays the output result of the convolutional neural network module and the data of the image storage module, but the function of the patent document is single, the system is incomplete, the algorithm accuracy is poor, and a processing mode for improving the accuracy is lacked.
The present invention has been made in view of this point.
Disclosure of Invention
The invention aims to provide an intelligent water-falling security system for solving the problems, when people who do not wear life jackets appear beside a shore or a wharf, the intelligent water-falling security system can alarm and remind related people, can accurately judge the people according to an optimized algorithm, can improve the accuracy of water-falling judgment and personnel information, can perform additional positioning assistance on the people falling into the water, and can improve the integrity of an alarm function.
The invention realizes the purpose through the following technical scheme, and the intelligent water-approaching security system comprises a detection module, a water falling identification module and a classification module, wherein the detection module, the water falling identification module and the classification module all need to operate on computer equipment, the detection module, the classification module and the water falling identification module utilize a YOLOv5 target detection algorithm to identify and detect personnel, perform Deepsort tracking on the personnel, calculate the number of the personnel by using a counter, and then record the personnel in a system platform, the detection module and the water falling identification module all need to be connected with a plurality of video acquisition devices, an alarm and an auxiliary positioning device are integrated on each video acquisition device, and a power supply connector is arranged on each video acquisition device through a power supply line.
Furthermore, in order to stably supply the number of pictures to ensure the operation of the security system, the video acquisition device is positioned at one side of the near-water area, and the video stream acquisition frame rate of the video acquisition device can acquire 30 frames per second and more than 30 frames per second.
Furthermore, in order to better limit the safety of the region close to water and exert a security function, in order to perform life jacket feature identification on each frame of picture acquired by the video acquisition device, after AlexNet classifies the result of the life jacket, only aiming at the current frame of result, combining depsort tracking, inputting each frame of information for prejudgment, and detecting the number of people in the current frame, the allowable number of people, namely the maximum allowable limit in the frame, by using a YOLOv5 detection network.
Further, in order to conveniently trigger an alarm and assist rescuers, the current picture is divided into a safe area and a water-falling area through a YOLOv5 detection network, life jacket characteristic identification is carried out on the safe area, life jacket characteristic identification is also carried out on the water-falling area, meanwhile, an auxiliary positioning device is activated to carry out matching detection, personnel information in the picture is conveniently managed, later-period images and information calling are carried out through a system platform, personnel characteristic information of the water-falling area is classified through an AlexNet algorithm, the position of a person is tracked by combining a Deepsort tracking algorithm, the Deepsort calculates and allocates an ID number of each person, and the person is bound and updated in real time.
Further, in order to reduce the processing amount and more optimally perform information judgment, in order to track the information of the bound people in each 20 frames of pictures, statistical lifejacket classification result statistics is performed on one person, if the number of lifejackets worn by the persons corresponding to each ID is larger than the number of the lifejackets not worn in the calculation results of the accumulated 20 frames, the persons corresponding to the ID are endowed with the lifejackets to wear for the next 20 frames, and the information is not refreshed in the next 20 frames until the next 20 frames are refreshed according to the statistical results.
Furthermore, the direction of the image acquisition is convenient to adjust, the protection performance of the video acquisition device is improved, the video acquisition device comprises a lifting frame, a rotating shell and a camera acquisition head, the lifting frame can adjust the heights of the camera acquisition head, the alarm and the auxiliary positioning device, so as to carry out image acquisition, alarm transmission and rescue assistance, a bearing seat is fixed on the lifting frame, a power motor is fixed at the bottom in the bearing seat, the bottom of the bearing seat is provided with air holes, the bottom of the rotating shell is provided with an air outlet pipe, a protective cover is fixed on the air outlet pipe, filter screens are fixed on the air outlet pipe and the inner walls of the air holes, an exhaust fan is fixed in the air outlet pipe, the alarm is fixed on the side wall of the bearing seat, the alarm comprises an alarm horn and a warning lamp, the warning lamp is fixed at the bottom of the alarm horn, the alarm horn is fixed on the side wall of the bearing seat through a support plate, a limit mechanism is arranged on the bearing seat, the limiting mechanism comprises a blocking annular wall, a bottom wheel and an inner roller, the blocking annular wall is fixed on the bearing seat, the bottom wheel is arranged at the bottom of the rotating shell, an extension plate is fixed on the inner wall of the rotating shell, a transmission rod is fixedly connected between an output shaft of the power motor and the extension plate, the inner roller is rotatably arranged at the bottom of the extension plate, the side wall of the inner roller is attached to the inner wall of the blocking ring wall, the rotating shell is fixed on the limiting mechanism, the camera collecting head is arranged in the rotating shell, a protective cover is fixed on the outer wall of the rotating shell, the protective cover is made of glass, and the protective cover is arranged on one side of the camera collecting head and one side of the infrared emitter.
Furthermore, in order to facilitate directional rescue of rescuers and reduce the time for finding people falling into water on site, the auxiliary positioning device comprises an infrared emitter and a laser emitter, the camera collecting head and the infrared emitter are mounted in the rotating shell through a power frame, the power frame comprises a fixed rod, a support plate and a driving motor, the support plate is fixed at one end of the fixed rod, a shaft rod is fixed on the side wall of the support plate, the driving motor is fixed on the side wall of the support plate, the shaft rod is connected with an output shaft of the driving motor, the camera collecting head and the infrared emitter are both fixedly mounted between the shaft rods, and the laser emitter is fixed on the infrared emitter.
The invention has the technical effects and advantages that:
the invention provides a near-water security monitoring and alarming system in all weather, which solves the problems of low accuracy, incomplete functions, easy occurrence of a large number of false alarms and the like in the prior art, and effectively improves the accuracy.
The invention can realize three main functions of drowning identification, life jacket identification and personnel overrun, a virtual boundary, namely a virtual fence is arranged between a safety region and a drowning region, and when personnel appear, whether the personnel have the life jacket or not can be analyzed, and safety warning can be sent out in time after the personnel do not have the life jacket and are overrun.
After people are detected by the drowning identification function through YOLOv5, tracking is carried out through a Deepsort algorithm, position information of each person is input, the Deepsort algorithm outputs the ID of each person, whether the drowning exceeds a certain threshold value is judged according to the longitudinal difference value between two frames of the positions of the pixel frames of the people corresponding to each ID between every two adjacent frames, if the drowning exceeds the certain threshold value, 10 frames of images before and after the drowning are sent to a TIN network for behavior identification reasoning, the drowning of the person is identified, the accuracy is greatly improved, and the false alarm rate is greatly reduced.
The video acquisition device can acquire pictures of an area near water, can rotate the acquisition direction according to actual requirements, can warn field personnel and convey information, and can assist in positioning the position of personnel falling into water by matching with the auxiliary positioning device, so that rescue personnel can conveniently perform rescue treatment.
Drawings
FIG. 1 is a functional flow diagram of the system of the present invention;
FIG. 2 is a flow chart of the system algorithm of the present invention;
fig. 3 is a flow chart of life jacket characteristic identification of the present invention;
FIG. 4 is a schematic control flow diagram of the present invention;
FIG. 5 is a schematic view of a video capture device according to the present invention;
FIG. 6 is a schematic view of the structure of FIG. 5 at A in accordance with the present invention;
FIG. 7 is a schematic cross-sectional view of a rotating housing of the present invention;
in the figure: 1. a video capture device; 101. a lifting frame; 102. rotating the shell; 103. a camera pick-up head; 2. an alarm; 3. an auxiliary positioning device; 301. an infrared emitter; 302. an infrared receiver; 4. a bearing seat; 5. and a limiting mechanism.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
Please refer to fig. 1-7, the intelligent security system for the water-facing, includes a detection module, a water falling identification module and a classification module, all of which need to operate on a computer device, the detection module, the water falling identification module and the classification module use YOLOv5 target detection algorithm to identify and detect people, perform deep sort tracking on people and use a counter to calculate the number of people, and then record in a system platform, all of which need to be connected with a video acquisition device 1, the video acquisition device 1 is provided with a plurality of devices, an alarm 2 and an auxiliary positioning device 3 are integrated on the video acquisition device 1, the video acquisition device 1 is located at one side of the water-facing area, the video acquisition device 1 acquires a video stream frame rate capable of acquiring more than 30 frames per second, and the video acquisition device 1 is provided with a power supply connector through a power supply line.
As a technical optimization scheme of the invention, life jacket feature recognition needs to be carried out on each frame of picture acquired by a video acquisition device 1, after AlexNet classification is carried out to obtain the result of the life jacket, only aiming at the result of the current frame, deepsort tracking is combined, each frame of information is input for prejudgment, and the number of people in the current picture, the allowed number of people, namely the allowed maximum limit amount in the picture, is 5 people through a YOLOv5 detection network.
As a technical optimization scheme of the invention, a current picture is subjected to region division through a YOLOv5 detection network and is divided into a safe region and a water falling region, life jacket feature recognition is carried out on the safe region, life jacket feature recognition is also carried out on the water falling region, meanwhile, an auxiliary positioning device 3 is activated to carry out matching detection, personnel information in the picture is managed conveniently, so that later-stage images and information retrieval are carried out through a system platform, personnel feature information of the water falling region is classified through an AlexNet algorithm, personnel are tracked through a Deepsort tracking algorithm, the Deepsort calculates and distributes ID numbers of each person, and the personnel are bound and updated in real time.
As a technical optimization scheme of the invention, in order to reduce the processing amount and more optimally judge information, in order to track the information of bound people, after every 20 frames of pictures, a statistical life jacket classification result statistics is carried out on one person, if the number of life jackets worn by the person corresponding to each ID is larger than the number of life jackets not worn in the calculation results of 20 frames of accumulated calculation results, the person corresponding to the ID is endowed with life jackets to wear in the next 20 frames, and the life jackets are not refreshed in the next 20 frames until the next 20 frames are refreshed according to the statistical result.
As a technical optimization scheme of the invention, the video acquisition device 1 comprises a lifting frame 101, a rotating shell 102 and a camera acquisition head 103, the camera acquisition head 103 can acquire a picture of visible light and a picture of an infrared light environment so as to improve the practicability, the lifting frame 101 can adjust the heights of the camera acquisition head 103, an alarm 2 and an auxiliary positioning device 3 so as to perform image acquisition, alarm transmission and rescue assistance, a bearing seat 4 is fixed on the lifting frame 101, a power motor is fixed at the bottom inside the bearing seat 4, vent holes are formed at the bottom of the bearing seat 4, an air outlet pipe is arranged at the bottom of the rotating shell 102, a protective cover is fixed on the air outlet pipe, filter screens are fixed on the inner walls of the air outlet pipe and the vent holes, an exhaust fan is fixed in the air outlet pipe, the alarm 2 is fixed on the side wall of the bearing seat 4, the alarm 2 comprises an alarm horn and a warning lamp, the warning lamp is fixed at the bottom of the warning horn, the warning horn is fixed on the side wall of the bearing seat 4 through the support plate, the bearing seat 4 is provided with the limiting mechanism 5, the limiting mechanism 5 comprises a blocking annular wall, a bottom wheel and an inner roller wheel, the blocking annular wall is fixed on the bearing seat 4, the bottom wheel is installed at the bottom of the rotating shell 102, the inner wall of the rotating shell 102 is fixed with an extension plate, a transmission rod is fixedly connected between an output shaft of the power motor and the extension plate, the inner roller wheel is rotatably installed at the bottom of the extension plate, the side wall of the inner roller wheel is attached to the inner wall of the blocking annular wall, the rotating shell 102 is fixed on the limiting mechanism 5, the camera acquisition head 103 is installed in the rotating shell 102, the outer wall of the rotating shell 102 is fixed with a protective cover made of glass, one side of the camera acquisition head 103 and the infrared emitter 301 is provided with a protective cover, when in use, the power motor can drive the rotating shell 102 to rotate, thereby change the picture collection direction of camera pick head 103, and when rotating shell 102 rotation, the bottom wheel can contact bears seat 4, avoids rotating shell 102's wearing and tearing, and the position that rotates shell 102 can be restricted to the gyro wheel in the cooperation, avoids rotating shell 102's rocking to block the rampart and can avoid outside rivers to get into, play the guard action.
As a technical optimization scheme of the invention, the auxiliary positioning device 3 comprises an infrared emitter 301 and a laser emitter 302, the infrared emitter 301 can emit an infrared beam, the camera collection head 103 and the infrared emitter 301 are both mounted in the rotating shell 102 through a power frame, the power frame comprises a fixed rod, a support plate and a driving motor, the support plate is fixed at one end of the fixed rod, a shaft rod is fixed on the side wall of the support plate, the driving motor is fixed on the side wall of the support plate, the shaft rod is connected with an output shaft of the driving motor, the camera collection head 103 and the infrared emitter 301 are both fixedly mounted between the shaft rods, the laser emitter 302 is fixed on the infrared emitter 301, the auxiliary positioning device 3 can perform auxiliary classification and auxiliary positioning on personnel falling into the water and rescue personnel, when the rescue personnel arrive, the infrared emitter 301 emits infrared light to perform auxiliary irradiation on the personnel falling into the water, then the camera collection head 103 can collect pictures, then the laser emitter 302 performs visible light irradiation on the area position of the person falling into the water, when the rescue personnel arrive, the rescue personnel can directly pass through a visible laser beam, the personnel can observe the person falling into the water quickly, the personnel can avoid drifting in the water, when the personnel also need to observe the infrared emitter 301, and the camera collection head can drive the camera emitter to rotate to drive the camera emitter 301 to rotate to drive the camera emitter to rotate to drive the camera emitter 301 to rotate to reach the appropriate angle.
When the device is used, the device is firstly connected with an external power supply through a power connector so as to supply power to a food collecting device, namely other electric appliances, when the device is used, a power motor can drive a rotating shell 102 to rotate, so that the picture collecting direction of a camera collecting head 103 is changed, when the rotating shell 102 rotates, a bottom wheel can contact a bearing seat 4 to avoid abrasion of the rotating shell 102, the position of the rotating shell 102 can be limited by matching with an inner roller, the swinging of the rotating shell 102 is avoided, the barrier ring wall can avoid external water flow from entering to play a role in protection, infrared rays emitted by an infrared ray emitter 301 can assist irradiation on people falling into water, then the camera collecting head 103 can collect infrared rays, visible light can be irradiated to the area position of people falling into water by a laser emitter 302, when rescue workers arrive, visible laser beams can be directly passed through to quickly observe the people falling into water, the people falling into water are prevented from drifting in the water, the rescue workers also need to observe and search, and when the irradiation angles of the camera collecting head 103 and the infrared ray emitter 301 need to be rotated to a proper angle, and the driving motor can drive the infrared ray emitter 301 or the camera collecting head 103 to rotate to a proper angle.
The detection module of the invention detects the personnel and the statistical personnel number of each frame through the video stream; the drowning identification module can firstly send the information of the detected personnel into a Deepsort algorithm for tracking and judging, and if the drowning exceeds a threshold value, send 20 frames before and after the frame into a TIN network for identifying the drowning; the classification module firstly sends the video stream to a network to classify whether a ship exists or not, then classifies the life jacket and improves the accuracy of life jacket classification by combining a tracking algorithm; the method utilizes a YOLOv5 target detection algorithm to detect people, then respectively carries out deep Sort tracking according to different functional requirements and uses a counter to calculate the number of people, so as to form three main functions of drowning identification, life jacket identification and personnel overrun, the YOLOv5 target detection algorithm is connected with an AlexNet classification algorithm in life jacket identification classification, after the people are detected by YOLOv5, the personnel coordinates of an original image are sent to an AlexNet network, the personnel images are obtained from an original image to be classified, whether the life jacket exists or not is obtained, after the result of the life jacket is obtained by AlexNet classification, only aiming at the result of a current frame, the deep Sort tracking is combined, each frame of information is input for prejudgment, and the result is statistically reported once by counting every 20 frames and refreshing once, so that the classification result of the life jacket is statistically reported, thereby greatly improving the classification accuracy in a real scene.
After people are detected by the drowning identification module through YOLOv5, tracking is carried out through a Deepsort algorithm, position information of each person is input, the Deepsort algorithm outputs the ID of each person, whether the drowning exceeds a certain threshold value is judged according to the longitudinal difference value between two frames of the positions of the pixel frames of the people corresponding to each ID between every two adjacent frames, if the drowning exceeds the certain threshold value, 10 frames of images before and after the drowning are sent to a TIN network for behavior identification reasoning, the drowning of the people is identified, the accuracy is greatly improved, and the false alarm rate is greatly reduced.
The method can realize intrusion alarm in a safety area, combines an AlexNet classification network through a YOLOv5 detection network, sets a pixel level safety range in a video picture, and alarms when people do not penetrate the life jacket into the safety range.
For the improvement of the classification accuracy of the life jacket, the classification accuracy of the AlexNet algorithm reaches 95%, and the classification accuracy of the life jacket is improved to 100% by combining with the Deepsort tracking algorithm. After people are detected by YOLOv5, each person is cut out and sent to an AlexNet network for classification, whether life jackets are worn or not is output, then position information of each person is input by combining a Deepsort tracking algorithm, the Deepsort calculates and distributes ID numbers of each person, then images of every 20 frames are calculated in an accumulated mode by tracking one person, statistical life jacket classification result statistics is carried out on one person, if the number of people wearing life jackets corresponding to each ID is larger than the number of people not wearing life jackets in the calculation result of 20 accumulated frames, the people corresponding to the ID are endowed with life jackets to wear in the next 20 frames, and the people do not refresh in the next 20 frames until the next 20 frames are refreshed according to the statistical result. In order to avoid tracking errors of the Deepsort algorithm, the Deepsort algorithm is updated continuously by adopting a method of refreshing 20 frames once. By adopting the method, the 95% accuracy of the existing algorithm can be improved to 100%.
The invention is used for solving the problems of low accuracy rate of life jacket detection and personnel overboard identification, imperfect function, easy occurrence of a large number of false alarms and the like in the existing method, improves the performance of the whole life jacket detection through the detection of personnel and a life jacket classification network including a tracking algorithm, and improves the performance of the overboard detection through the detection and tracking of personnel and the behavior identification of human body movement, thereby effectively improving the accuracy rate of system functions.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (10)
1. Face water intelligent security system, including detection module, fall water identification module and classification module, its characterized in that: the detection module, the identification module that falls into water and the classification module all need operate on computer equipment, detection module with the identification module that falls into water all needs connect video acquisition device (1), video acquisition device (1) is equipped with a plurality ofly, and integrated on video acquisition device (1) has siren (2) and auxiliary positioning device (3), be equipped with power connection through the power cord on video acquisition device (1).
2. The intelligent security system near water of claim 1, characterized in that: the video acquisition device (1) is located on one side of a waterside area, and the video acquisition device (1) can acquire 30 frames of pictures per second or more.
3. The intelligent security system in presence of water of claim 2, characterized in that: the detection module, the classification module and the drowning identification module carry out identification detection on the personnel by utilizing a YOLOv5 target detection algorithm, carry out Deepsort tracking on the personnel, use a counter to calculate the number of the personnel, and then record the personnel in a system platform.
4. The intelligent security system near water of claim 3, characterized in that: life jacket feature recognition needs to be carried out on each frame of picture acquired by the video acquisition device (1), and after the result of the life jacket is obtained by AlexNet classification, each frame of information is input for prejudgment only by combining Deepsort tracking with the result of the current frame.
5. The intelligent security system in presence of water of claim 4, characterized in that: the number of people in the current picture is detected through a YOLOv5 detection network, and the allowed number of people, namely the maximum allowable limit amount in the picture is 5 people.
6. The intelligent security system in presence of water of claim 5, characterized in that: the method comprises the steps of carrying out area division on a current picture through a YOLOv5 detection network, dividing the current picture into a safe area and a water falling area, carrying out life jacket feature recognition on the safe area, carrying out life jacket feature recognition on the water falling area, and activating an auxiliary positioning device (3) to carry out matching detection.
7. The intelligent security system in presence of water of claim 6, characterized in that: the personnel characteristic information of the water falling area is classified through an AlexNet algorithm, the position of the personnel is tracked by combining a Deepsort tracking algorithm, the Deepsort calculates and distributes the ID number of each person, and the personnel is bound and updated in real time.
8. The intelligent security system near water of claim 7, characterized in that: and tracking the information of the bound personnel, counting the statistical life jacket classification result of one person after every 20 frames of pictures, if the number of life jackets worn by the person corresponding to each ID is larger than the number of life jackets not worn in the calculation results of the accumulated 20 frames, giving the life jacket to the person corresponding to the ID for the next 20 frames, and not refreshing the person in the next 20 frames until the next 20 frames are refreshed according to the statistical result.
9. The intelligent security system in presence of water of claim 1, characterized in that: the video acquisition device (1) comprises a lifting frame (101), a rotating shell (102) and a camera acquisition head (103), wherein a bearing seat (4) is fixed on the lifting frame (101), an alarm (2) is fixed on the side wall of the bearing seat (4), a limiting mechanism (5) is installed on the bearing seat (4), the rotating shell (102) is fixed on the limiting mechanism (5), and the camera acquisition head (103) is installed in the rotating shell (102).
10. The intelligent security system near water of claim 9, characterized in that: the auxiliary positioning device (3) comprises an infrared emitter (301) and a laser emitter (302), the camera shooting collecting head (103) and the infrared emitter (301) are both installed in the rotating shell (102) through a power frame, and the laser emitter (302) is fixed on the infrared emitter (301).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210811561.4A CN115311820A (en) | 2022-07-11 | 2022-07-11 | Intelligent security system near water |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210811561.4A CN115311820A (en) | 2022-07-11 | 2022-07-11 | Intelligent security system near water |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115311820A true CN115311820A (en) | 2022-11-08 |
Family
ID=83856392
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210811561.4A Pending CN115311820A (en) | 2022-07-11 | 2022-07-11 | Intelligent security system near water |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115311820A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116778673A (en) * | 2023-08-17 | 2023-09-19 | 瞳见科技有限公司 | Water area safety monitoring method, system, terminal and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111028480A (en) * | 2019-12-06 | 2020-04-17 | 江西洪都航空工业集团有限责任公司 | Drowning detection and alarm system |
| CN111540173A (en) * | 2020-04-22 | 2020-08-14 | 上海振华重工电气有限公司 | Drowning alarm system and method based on intelligent picture recognition |
| WO2021114702A1 (en) * | 2019-12-10 | 2021-06-17 | 中国银联股份有限公司 | Target tracking method, apparatus and system, and computer-readable storage medium |
| CN113271438A (en) * | 2021-05-13 | 2021-08-17 | 烟台文化旅游职业学院 | Safety supervision system for bathing beach |
| CN114023003A (en) * | 2021-10-26 | 2022-02-08 | 安徽唐后智能科技有限公司 | Intelligent shoreside drowning rescue equipment |
| CN114633855A (en) * | 2022-04-01 | 2022-06-17 | 浙江机电职业技术学院 | Water area safety early warning device and control method thereof |
| CN114674323A (en) * | 2022-04-12 | 2022-06-28 | 北京邮电大学 | Intelligent indoor navigation method based on image target detection and tracking |
-
2022
- 2022-07-11 CN CN202210811561.4A patent/CN115311820A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111028480A (en) * | 2019-12-06 | 2020-04-17 | 江西洪都航空工业集团有限责任公司 | Drowning detection and alarm system |
| WO2021114702A1 (en) * | 2019-12-10 | 2021-06-17 | 中国银联股份有限公司 | Target tracking method, apparatus and system, and computer-readable storage medium |
| CN111540173A (en) * | 2020-04-22 | 2020-08-14 | 上海振华重工电气有限公司 | Drowning alarm system and method based on intelligent picture recognition |
| CN113271438A (en) * | 2021-05-13 | 2021-08-17 | 烟台文化旅游职业学院 | Safety supervision system for bathing beach |
| CN114023003A (en) * | 2021-10-26 | 2022-02-08 | 安徽唐后智能科技有限公司 | Intelligent shoreside drowning rescue equipment |
| CN114633855A (en) * | 2022-04-01 | 2022-06-17 | 浙江机电职业技术学院 | Water area safety early warning device and control method thereof |
| CN114674323A (en) * | 2022-04-12 | 2022-06-28 | 北京邮电大学 | Intelligent indoor navigation method based on image target detection and tracking |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116778673A (en) * | 2023-08-17 | 2023-09-19 | 瞳见科技有限公司 | Water area safety monitoring method, system, terminal and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111770266B (en) | Intelligent visual perception system | |
| CN101188745B (en) | Intelligent drowning video monitoring system and method for natatorium | |
| CN206400717U (en) | A kind of security protection patrol robot based on GPS | |
| CN106251701A (en) | Based on the vehicle rearview monitor and alarm system and the method that rotate zoom multi-cam | |
| CN108227606A (en) | A kind of ship security intelligent management system perceived based on multi-source | |
| CN210899299U (en) | Tunnel monitoring system | |
| WO2022142973A1 (en) | Robot protection system and method | |
| CN107662867A (en) | The step roller monitoring of passenger conveyor and maintenance operation monitored by personnel | |
| CN109996037A (en) | A kind of method for safety monitoring in waters, device, equipment and system | |
| CN114566056A (en) | Highway tunnel driving safety risk identification, prevention and control method and system | |
| JP2014194686A (en) | Crime prevention system | |
| CN105448017A (en) | Community rail alarming system | |
| CN115311820A (en) | Intelligent security system near water | |
| CN111523386B (en) | High-speed railway platform door monitoring and protecting method and system based on machine vision | |
| CN112263803A (en) | Unmanned vehicle intelligent security system based on real-time scene inspection and automatic detection fire extinguishing and control method | |
| CN115841730A (en) | Video monitoring system and abnormal event detection method | |
| KR101415848B1 (en) | Monitering apparatus of school-zone using detection of human body and vehicle | |
| CN118397571B (en) | Reservoir intelligent monitoring method and system based on AI video technology | |
| CN114155470A (en) | River channel area intrusion detection method, system and storage medium | |
| CN110222216A (en) | Multifunctional intellectual defence system | |
| KR102580549B1 (en) | Safety Information Broadcasting System Using Artificial Intelligence | |
| CN111063148A (en) | Remote night vision target video detection method | |
| CN116494241A (en) | Robot security control system | |
| KR102686900B1 (en) | Processing apparatus for dangerous situation information using smart pole | |
| CN116797967A (en) | Visual monitoring hidden trouble identification method and device for overhead transmission line |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221108 |
|
| RJ01 | Rejection of invention patent application after publication |