CN110751063A - Infant quilt kicking prevention recognition device and method based on deep learning - Google Patents
Infant quilt kicking prevention recognition device and method based on deep learning Download PDFInfo
- Publication number
- CN110751063A CN110751063A CN201910937987.2A CN201910937987A CN110751063A CN 110751063 A CN110751063 A CN 110751063A CN 201910937987 A CN201910937987 A CN 201910937987A CN 110751063 A CN110751063 A CN 110751063A
- Authority
- CN
- China
- Prior art keywords
- infant
- quilt
- deep learning
- image
- information
- 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
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000013135 deep learning Methods 0.000 title claims abstract description 25
- 230000002265 prevention Effects 0.000 title claims description 12
- 230000006399 behavior Effects 0.000 claims abstract description 12
- 238000001514 detection method Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 14
- 230000003068 static effect Effects 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 238000004806 packaging method and process Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 238000005516 engineering process Methods 0.000 claims description 2
- 238000013473 artificial intelligence Methods 0.000 description 6
- 230000009471 action Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 230000007958 sleep Effects 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 238000003825 pressing Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000001931 thermography Methods 0.000 description 2
- 238000012356 Product development Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000036760 body temperature Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
Images
Classifications
-
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/29—Graphical models, e.g. Bayesian networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/103—Static body considered as a whole, e.g. static pedestrian or occupant recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a device and a method for identifying a quilt kicked by an infant based on deep learning, and relates to the technical field of intelligent control. The method identifies the information of the face, the skeleton, the bed, the quilt and the like of the infant in the acquired image to be identified through a deep learning algorithm, and calculates whether the infant kicks off the quilt or not according to the position of the face of the infant, the information number of the skeleton of the infant, the position of the infant and the position of the bed. After the quilt kicking behavior is found, the guardian is reminded through the sound playing equipment, so that the alarm can be given in time.
Description
Technical Field
The invention relates to the technical field of intelligent control, in particular to a quilt kicking prevention identification device and method for infants based on deep learning.
Background
The development of the current artificial intelligence industry is a wave, mainly comes from the proposal of a deep learning algorithm, realizes large-scale calculation on the basis of data volume and calculation capacity, and belongs to technical breakthrough. Artificial intelligence has been widely used in various fields such as internet and mobile internet applications, automatic driving, shared travel, intelligent finance, service robots, intelligent manufacturing, artificial intelligence assisted education, smart agriculture, and intelligent customer service. In the field of home administration service, artificial intelligence is widely applied, but the artificial intelligence is not applied to solve the problem that a quilt is kicked by an infant. At present, the infant kicking-proof devices in the market can be roughly divided into the following types:
1. the physical model device is used for preventing the infant from kicking the quilt, for example: the quilt is fixed by the bottom plate and the pressing block, the quilt is tightly pressed between the convex ribs and the grooves after the convex ribs on the pressing block are matched with the grooves on the bottom plate, and the pressing block is buckled on the bottom plate through the lock catch structure, so that the quilt can be fixed on the quilt kicking prevention structure, and the quilt is prevented from being kicked by infants. The device has the advantages of simple structure and low cost, but the mode is troublesome and not intelligent, and can only play a certain preventive role without subsequent action.
2. The infant kicking quilt is monitored by means of thermal imaging and an infrared sensor, but the method is easily influenced by a shelter, a light source and the like to cause misjudgment, and is low in accuracy and high in price.
3. The body temperature of the infant is monitored by a wearable temperature monitor. This mode need be at children's pajamas or install a plurality of temperature sensor on one's body, can't avoid reducing the travelling comfort of sleep, can direct influence sleep even. In addition, the cooling speed of the body surface or the pajamas is too slow, and the variation range is small, so that the alarm cannot be given in time in the mode, and the risk of catching a cold still exists. Meanwhile, the method has high requirements on the sensitivity of the temperature sensor, and the cost is increased.
4. By attaching the retaining contacts to the corners. When the child kicks off the quilt, the limiting contactor is pulled to conduct the alarm circuit to give an alarm. However, this method cannot avoid that the child may pull the limit contactor when turning over, resulting in false alarm.
Disclosure of Invention
The invention aims to: the invention provides a device and a method for identifying an infant quilt kicking prevention based on deep learning, aiming at solving the problems of untimely and wrong alarming in the existing solving mode that most children have the action of kicking the quilt but can not actively cover the quilt in the sleeping process, so that the children are easy to catch a cold.
The technical scheme adopted by the invention is as follows:
the utility model provides an infant prevents playing quilt recognition device based on degree of depth study, includes the power supply unit who provides electric energy for each component of device, still includes:
the video acquisition equipment is used for acquiring the sleeping state information of the infant and sending the acquired information to the core processor;
the core processor is internally provided with a GPU server and is used for judging whether the quilt kicking condition exists in the infant or not according to the information acquired by the video acquisition equipment;
and the sound playing equipment is used for calling the guardian when the core processor identifies that the child kicks the quilt.
Further, the baby bed also comprises a video display device for displaying the sleeping condition of the baby.
Furthermore, the device also comprises a mounting box which is used for packaging all components of the device and fixing the whole device.
Further, the power supply equipment adopts an external 12V power supply; the video acquisition equipment adopts a camera with a dimming technology, and the camera can normally image in the daytime and at night; the video display equipment adopts a five-inch display screen.
The identification method of the identification device for preventing the child from kicking off the quilt comprises the following steps:
step 1: acquiring a large number of sample pictures of the kicking behavior of the infant as sample data by using video acquisition equipment, training the sample data by adopting a face detection algorithm of YOLO v3, generating weights and weight files for identifying the infant, a bed and a quilt, and updating the weights in the face detection algorithm of YOLO v3 according to the generated weights;
step 2: preprocessing the daytime and nighttime pictures of the sample picture by utilizing an OpenCV image processing algorithm to finish positioning the infant on the bed;
and step 3: packaging the improved face detection algorithm of YOLO v3, the OpenCV image processing algorithm, the human body posture estimation algorithm based on OpenPose and the weight file generated by training into a local deep learning algorithm and integrating the local deep learning algorithm into a core processor;
and 4, step 4: the video acquisition equipment acquires images once at intervals of a certain time, transmits acquired image data to be identified to the core processor, identifies the sleeping state information of the infant through a local deep learning algorithm, judges whether the infant kicks off a quilt or not according to the sleeping state information of the infant, and calls a guardian through the sound playing equipment if the quilt is kicked off.
Furthermore, the infant sleeping state information includes an infant position, a bed position, an infant face position, and a skeleton information number.
Further, the specific steps of identifying the sleeping state information of the infant through the local deep learning algorithm are as follows:
step A: the method comprises the steps that image data to be recognized are collected once every fixed time by a video collection device, digital image comparison is carried out on N adjacent photos in a time domain space through an OpenCV image processing algorithm, images of a target object at a relatively static state moment are obtained, image histogram information is read, and a primary image is obtained through balance processing;
and B: reading the primary image, carrying out infant target positioning on the obtained photo by using an improved face detection algorithm of YOLO v3, realizing target identification and infant marking of a plurality of infant targets in a target area, and judging the positions and sizes of the infants; if no infant object is found, the device is rested for a certain time, and the step A is repeated; if the infant object is found, acquiring and storing an image of the infant area, recording the image as a secondary image, and executing the step C;
and C: extracting information of joint points by using a human body posture estimation algorithm based on OpenPose, fusing a posture network and a skeleton network, identifying skeleton position information of the infant, and judging whether the infant kicks off a quilt according to the condition that each skeleton of the infant is covered; after judging that the infant kicks off the quilt, starting counting, if the counting times reach M times, the core processor immediately sends an instruction to the sound playing equipment, so that the sound playing equipment sends out a warning sound that the infant kicks off the quilt: and if the counting times do not reach M times, continuously repeating the steps A to C.
In summary, due to the adoption of the technical scheme, the invention has the beneficial effects that:
1. according to the invention, through image preprocessing, photos in the day and at night can be correctly read, information such as the face of the infant, the skeleton of the infant, the bed, the quilt and the like is identified through a deep learning algorithm, and whether the infant kicks away the quilt or not is calculated according to the position of the face of the infant, the information number of the skeleton of the infant, the position of the infant and the position of the bed. After the quilt kicking behavior is found, the guardian is reminded through the sound playing equipment, so that the alarm is given in time, and the practicability is high.
2. In the invention, the device is started, and the video information is acquired by the video acquisition equipment and displayed on the video display equipment, so that the position of the device is convenient to adjust. After the adjustment is finished, the device is in a static state, and a user can independently select whether to close the video display equipment. If the user selects to start the video display device, the user can observe the sleeping condition of the infant through the video display device at any time; if the video display equipment is selected to be turned off, the power consumption cost is saved, meanwhile, the personal privacy is prevented from being revealed, and the safety is high.
3. According to the device and the method, artificial intelligence plays a positive role in the field of housekeeping services, the product development in the field is obviously promoted, and particularly, an especially important step is taken in the research of preventing the infant from kicking the quilt. The invention can be applied to individual families and a large number of infant sleeping sites in kindergartens, and has high practicability.
4. Compared with the traditional physical model quilt kicking prevention device, the invention improves the defects of prevention only and no treatment measures; compared with a sensor and a quilt kicking prevention device for thermal imaging, the invention improves the defect that misjudgment is easily caused by the influence of external factors such as shelters, light sources and the like; compared with the wearable temperature monitor device, the invention improves the defect that the alarm cannot be given in time. The invention can realize intelligent, automatic and accurate detection in real sense, and alarm in time, so that the behavior of kicking the quilt by the infant during sleep can be prevented best and alarm in time.
5. In the invention, the improved face detection algorithm of the YOLO v3 shows excellent resistance to rotation, scale change, face recognition and shielding, so that the device can quickly, accurately and stably carry out continuous detection and recognition on infants.
6. In the invention, the human body posture estimation algorithm based on OpenPose fuses the posture network and the skeleton network based on the graph convolution network model of the skeleton information, thereby further improving the accuracy of behavior recognition and reducing the recognition error rate of the human body posture estimation algorithm on the skeleton information of the infant.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.
FIG. 1 is a block diagram of the apparatus of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the detailed description and specific examples, while indicating the preferred embodiment of the invention, are intended for purposes of illustration only and are not intended to limit the scope of the invention. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
It is noted that relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. The term "comprising", without further limitation, means that the element so defined is not excluded from the group consisting of additional identical elements in the process, method, article, or apparatus that comprises the element.
The features and properties of the present invention are described in further detail below with reference to examples.
Example one
In a preferred embodiment of the present invention, there is provided a deep learning-based quilt kicking prevention recognition device for infants, as shown in fig. 1, the device comprising:
the power supply equipment is used for supplying electric energy to all components of the device;
the video acquisition equipment is used for acquiring the sleeping state information of the infant and sending the acquired information to the core processor;
the core processor is internally provided with a GPU server and is used for judging whether the quilt kicking condition exists in the infant or not according to the information acquired by the video acquisition equipment;
and the sound playing equipment is used for calling the guardian when the core processor identifies that the child kicks the quilt.
All the components of the device are packaged through the mounting box and the whole device is fixed, so that a fully-automatic, systematic and intelligent child quilt kicking prevention device is formed. In this embodiment, power equipment adopts external 12V power, and video acquisition equipment adopts the infrared camera that has the technique of adjusting luminance, and infrared camera is daytime and night all can normally form images, and video display equipment adopts five cun display screens. When the child kicks off the quilt for various reasons, the device automatically sends out warning information to the guardian through intelligent judgment, thereby preventing the child from getting ill due to the fact that the child kicks off the quilt and catches a cold.
The identification method adopting the identification device for preventing the quilt from being kicked by the infant comprises the following steps:
step 1: a video acquisition device is used for acquiring a large number of sample pictures of the kicking behavior of the infant as sample data, the sample data are trained by adopting the face detection algorithm of YOLO v3, weights and weight files for identifying the infant, the bed and the quilt are generated, and the weights in the face detection algorithm of YOLO v3 are updated according to the generated weights.
Step 2: the OpenCV image processing algorithm is utilized to carry out pretreatment on the daytime and nighttime images of the sample image, so that the image data to be identified (generated in the step 4) can be guaranteed to be capable of effectively shooting the outline of the infant, and the positioning of the infant on the bed can be completed.
And step 3: the improved face detection algorithm of YOLO v3, the OpenCV image processing algorithm, the human body posture estimation algorithm based on OpenPose and the weight file generated by training are packaged into a local deep learning algorithm and integrated into a core processor.
The human body posture estimation algorithm based on OpenPose utilizes a recurrent neural network to extract the characteristic information of human body joint points, and adds the understanding of time domain information. According to the algorithm, the posture network and the skeleton network are fused by researching a graph convolution network model based on skeleton information, so that the accuracy of behavior recognition is further improved, and the algorithm is used for recognizing the skeleton information of the infant.
And 4, step 4: the video acquisition equipment acquires images once at intervals of a certain time, transmits acquired image data to be identified to the core processor, identifies the sleeping state information of the infant through a local deep learning algorithm, judges whether the infant kicks off a quilt or not according to the sleeping state information of the infant, and calls a guardian through the sound playing equipment if the quilt is kicked off. The infant sleeping state information comprises an infant position, a bed position, an infant face position and a skeleton information number.
Specifically, image data to be recognized acquired from video acquisition equipment is subjected to image preprocessing through a local deep learning algorithm, information such as the face of an infant, the skeleton of the infant, a bed and a quilt is recognized, whether the infant kicks away the quilt is calculated through the position of the infant, the position of the bed, the position of the face of the infant and skeleton information numbering, whether the infant, the bed and the quilt exist is recognized through face detection, whether the infant is shielded by the quilt is judged through the human skeleton detected through a human posture estimation algorithm based on OpenPose, and therefore the judgment of whether the infant kicks away the quilt is made. If quilt kicking behaviors are found, music is made to remind a guardian through the sound playing device in time, and the design meets the requirement of processing the response time of the infant for kicking the quilt.
The specific steps of the analysis of the quilt kicking behavior of the infant in the step are as follows:
step A: the method comprises the steps that image data to be recognized are collected once every 2s by video collection equipment, digital image comparison is carried out on a plurality of adjacent pictures in a time domain space through an OpenCV image processing algorithm, images of a target object at a relatively static state moment are obtained, image histogram information is read, and parameters such as image contrast and brightness are adjusted through equalization processing, so that image outline information is clearer. The image information is then saved in a specific location for further use.
And B: reading the image of the previous step, and carrying out infant target positioning on the obtained picture by utilizing the quick positioning characteristic of the improved face detection algorithm of YOLO v3, so as to realize the target identification and infant marking of a plurality of infant targets in the target area and judge the positions and sizes of the infants. If the infant object is not found in the current step, the subsequent steps are not carried out, the system takes a rest for 5 seconds, and the first step is continuously repeated. If a child object is found, the child area image is captured and saved to provide image data for the next step.
And C: when the current step is reached, the image positioning is basically accurate and very specific, the information of joint points is extracted by a human body posture estimation algorithm based on OpenPose, a posture network is fused with a skeleton network, the skeleton position information of the infant is identified, according to the display condition of skeleton numbers of the infant, such as legs, arms, heads, bodies and the like, whether the infant kicks off a quilt or not is judged according to the position where each skeleton of the infant is covered, namely if the trunk part and the legs of the infant are identified, the infant is judged to kick off the quilt, and the skeleton direction position of the infant can also record the sleeping posture direction of the infant. The step can also record the current image of the quilt kicked open by the infant and mark the sleeping state of the infant as the quilt kicked open.
And after judging that the infant kicks off the quilt, counting, if the counting times reach 5 times, indicating that the continuous 5-time image acquisition is in a state of kicking off the quilt, and at the moment, immediately sending an instruction to the sound playing equipment by the core processor to enable the sound playing equipment to send out a warning sound for the infant to kick off the quilt. And if the times do not reach 5, continuously repeating the steps A to C. The method can be used for processing the phenomenon that the quilt kicking behavior of the infant occurs within 1 minute when the infant sleeps under the complex background condition, so that the intelligent judgment of the quilt kicking behavior is realized, and early warning is given out.
Example two
The embodiment further includes a video display device for displaying the sleeping status of the infant on the basis of the first embodiment. And starting the device, acquiring video information through the video acquisition equipment, and displaying the video information on the video display equipment, so that the position of the device can be conveniently adjusted. After the adjustment is finished, the device is in a static state, and a user can independently select whether to close the video display equipment. If the user selects to start the video display device, the user can observe the sleeping condition of the infant through the video display device at any time. If the video display equipment is selected to be turned off, the power consumption cost is saved, meanwhile, the personal privacy is prevented from being revealed, and the safety is high.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (7)
1. The utility model provides an infant prevents playing quilt recognition device based on degree of depth study, includes the power supply unit who provides electric energy for each component part of device, its characterized in that still includes:
the video acquisition equipment is used for acquiring the sleeping state information of the infant and sending the acquired information to the core processor;
the core processor is internally provided with a GPU server and is used for judging whether the quilt kicking condition exists in the infant or not according to the information acquired by the video acquisition equipment;
and the sound playing equipment is used for calling the guardian when the core processor identifies that the child kicks the quilt.
2. The child anti-kicking quilt recognition device based on deep learning of claim 1, further comprising a video display device for displaying the sleeping status of the child.
3. The quilt recognition device for preventing children from kicking off based on deep learning of claim 1, further comprising a mounting box for packaging the components of the device and fixing the whole device.
4. The infant quilt kicking prevention recognition device based on deep learning of claim 1 or 2, wherein the power supply equipment adopts an external 12V power supply; the video acquisition equipment adopts a camera with a dimming technology, and the camera can normally image in the daytime and at night; the video display equipment adopts a five-inch display screen.
5. The identification method of the child quilt kicking prevention identification device of any one of claims 1 to 4 is characterized by comprising the following steps:
step 1: acquiring a large number of sample pictures of the kicking behavior of the infant as sample data by using video acquisition equipment, training the sample data by adopting a face detection algorithm of YOLO v3, generating weights and weight files for identifying the infant, a bed and a quilt, and updating the weights in the face detection algorithm of YOLO v3 according to the generated weights;
step 2: preprocessing the daytime and nighttime pictures of the sample picture by utilizing an OpenCV image processing algorithm to finish positioning the infant on the bed;
and step 3: packaging the improved face detection algorithm of YOLO v3, the OpenCV image processing algorithm, the human body posture estimation algorithm based on OpenPose and the weight file generated by training into a local deep learning algorithm and integrating the local deep learning algorithm into a core processor;
and 4, step 4: the video acquisition equipment acquires images once at intervals of a certain time, transmits acquired image data to be identified to the core processor, identifies the sleeping state information of the infant through a local deep learning algorithm, judges whether the infant kicks off a quilt or not according to the sleeping state information of the infant, and calls a guardian through the sound playing equipment if the quilt is kicked off.
6. The method for identifying a quilt kicked off by an infant based on deep learning as claimed in claim 5, wherein the sleeping state information of the infant comprises the position of the infant, the position of a bed, the position of the face of the infant and a skeleton information number.
7. The method for identifying the quilt kicked off by the infant based on the deep learning as claimed in claim 5, wherein the specific steps of identifying the sleeping state information of the infant through the local deep learning algorithm are as follows:
step A: the method comprises the steps that image data to be recognized are collected once every fixed time by a video collection device, digital image comparison is carried out on N adjacent photos in a time domain space through an OpenCV image processing algorithm, images of a target object at a relatively static state moment are obtained, image histogram information is read, and a primary image is obtained through balance processing;
and B: reading the primary image, carrying out infant target positioning on the obtained photo by using an improved face detection algorithm of YOLO v3, realizing target identification and infant marking of a plurality of infant targets in a target area, and judging the positions and sizes of the infants; if no infant object is found, the device is rested for a certain time, and the step A is repeated; if the infant object is found, acquiring and storing an image of the infant area, recording the image as a secondary image, and executing the step C;
and C: extracting information of joint points by using a human body posture estimation algorithm based on OpenPose, fusing a posture network and a skeleton network, identifying skeleton position information of the infant, and judging whether the infant kicks off a quilt according to the condition that each skeleton of the infant is covered; after judging that the infant kicks off the quilt, starting counting, and if the counting times reach M times, immediately sending an instruction to the sound playing equipment by the core processor to enable the sound playing equipment to send out a warning sound for the infant to kick off the quilt; and if the counting times do not reach M times, continuously repeating the steps A to C.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910937987.2A CN110751063A (en) | 2019-09-29 | 2019-09-29 | Infant quilt kicking prevention recognition device and method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910937987.2A CN110751063A (en) | 2019-09-29 | 2019-09-29 | Infant quilt kicking prevention recognition device and method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110751063A true CN110751063A (en) | 2020-02-04 |
Family
ID=69277488
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910937987.2A Pending CN110751063A (en) | 2019-09-29 | 2019-09-29 | Infant quilt kicking prevention recognition device and method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110751063A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111460978A (en) * | 2020-03-30 | 2020-07-28 | 中国科学院自动化研究所南京人工智能芯片创新研究院 | Infant behavior monitoring system based on motion judgment sensor and deep learning technology and judgment method thereof |
| CN112001347A (en) * | 2020-08-31 | 2020-11-27 | 重庆科技学院 | Motion recognition method based on human skeleton shape and detection target |
| CN112712020A (en) * | 2020-12-29 | 2021-04-27 | 文思海辉智科科技有限公司 | Sleep monitoring method, device and system |
| CN112874463A (en) * | 2021-03-01 | 2021-06-01 | 长安大学 | Protection and alarm system and method for children trapped in high-temperature vehicle |
| CN113221802A (en) * | 2021-05-24 | 2021-08-06 | 新疆爱华盈通信息技术有限公司 | Quilt kicking identification method and device and electronic equipment |
| CN113343891A (en) * | 2021-06-24 | 2021-09-03 | 深圳市起点人工智能科技有限公司 | Detection device and detection method for child kicking quilt |
| CN113465155A (en) * | 2021-06-23 | 2021-10-01 | 深圳市海清视讯科技有限公司 | Method, system, host and storage medium for reducing quilt kicking behavior of infants |
| CN114067435A (en) * | 2021-11-15 | 2022-02-18 | 山东大学 | Sleep behavior detection method and system based on pseudo-3D convolutional network and attention mechanism |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102789672A (en) * | 2012-07-03 | 2012-11-21 | 北京大学深圳研究生院 | Intelligent identification method and device for baby sleep position |
| CN105096527A (en) * | 2015-06-04 | 2015-11-25 | 上海卓易科技股份有限公司 | Monitoring and alerting method and system based on child's sleep state |
| CN108229421A (en) * | 2018-01-24 | 2018-06-29 | 华中科技大学 | A kind of falling from bed behavior real-time detection method based on deep video information |
| CN108615333A (en) * | 2018-05-25 | 2018-10-02 | 朱利 | Infant asphyxia early warning system based on artificial intelligence and method |
| CN109840493A (en) * | 2019-01-27 | 2019-06-04 | 武汉星巡智能科技有限公司 | Infantal sleeping condition detection method, device and computer readable storage medium |
| CN109886137A (en) * | 2019-01-27 | 2019-06-14 | 武汉星巡智能科技有限公司 | Infant sleeping posture detection method, device and computer readable storage medium |
| CN110008816A (en) * | 2019-01-28 | 2019-07-12 | 温州大学 | A kind of method that real-time detection baby kicks quilt son |
-
2019
- 2019-09-29 CN CN201910937987.2A patent/CN110751063A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102789672A (en) * | 2012-07-03 | 2012-11-21 | 北京大学深圳研究生院 | Intelligent identification method and device for baby sleep position |
| CN105096527A (en) * | 2015-06-04 | 2015-11-25 | 上海卓易科技股份有限公司 | Monitoring and alerting method and system based on child's sleep state |
| CN108229421A (en) * | 2018-01-24 | 2018-06-29 | 华中科技大学 | A kind of falling from bed behavior real-time detection method based on deep video information |
| CN108615333A (en) * | 2018-05-25 | 2018-10-02 | 朱利 | Infant asphyxia early warning system based on artificial intelligence and method |
| CN109840493A (en) * | 2019-01-27 | 2019-06-04 | 武汉星巡智能科技有限公司 | Infantal sleeping condition detection method, device and computer readable storage medium |
| CN109886137A (en) * | 2019-01-27 | 2019-06-14 | 武汉星巡智能科技有限公司 | Infant sleeping posture detection method, device and computer readable storage medium |
| CN110008816A (en) * | 2019-01-28 | 2019-07-12 | 温州大学 | A kind of method that real-time detection baby kicks quilt son |
Non-Patent Citations (2)
| Title |
|---|
| 杨路明主编: "《C语言程序设计教程》", 30 November 2018, 北京邮电大学出版社 * |
| 裴明涛主编: "《视频事件的分析与理解》", 31 March 2019, 北京理工大学出版社 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111460978A (en) * | 2020-03-30 | 2020-07-28 | 中国科学院自动化研究所南京人工智能芯片创新研究院 | Infant behavior monitoring system based on motion judgment sensor and deep learning technology and judgment method thereof |
| CN112001347A (en) * | 2020-08-31 | 2020-11-27 | 重庆科技学院 | Motion recognition method based on human skeleton shape and detection target |
| CN112712020A (en) * | 2020-12-29 | 2021-04-27 | 文思海辉智科科技有限公司 | Sleep monitoring method, device and system |
| CN112712020B (en) * | 2020-12-29 | 2024-05-31 | 文思海辉智科科技有限公司 | Sleep monitoring method, device and system |
| CN112874463A (en) * | 2021-03-01 | 2021-06-01 | 长安大学 | Protection and alarm system and method for children trapped in high-temperature vehicle |
| CN113221802A (en) * | 2021-05-24 | 2021-08-06 | 新疆爱华盈通信息技术有限公司 | Quilt kicking identification method and device and electronic equipment |
| CN113465155A (en) * | 2021-06-23 | 2021-10-01 | 深圳市海清视讯科技有限公司 | Method, system, host and storage medium for reducing quilt kicking behavior of infants |
| CN113343891A (en) * | 2021-06-24 | 2021-09-03 | 深圳市起点人工智能科技有限公司 | Detection device and detection method for child kicking quilt |
| CN114067435A (en) * | 2021-11-15 | 2022-02-18 | 山东大学 | Sleep behavior detection method and system based on pseudo-3D convolutional network and attention mechanism |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110751063A (en) | Infant quilt kicking prevention recognition device and method based on deep learning | |
| Lu et al. | Deep learning for fall detection: Three-dimensional CNN combined with LSTM on video kinematic data | |
| Wang et al. | Toward a passive low-cost in-home gait assessment system for older adults | |
| CN107609517B (en) | Classroom behavior detection system based on computer vision | |
| CN106464844A (en) | Systems and methods for configuring baby monitor cameras to provide uniform data sets for analysis | |
| CN102789672B (en) | A kind of baby's sleeping position intelligent identification Method and device | |
| CN110196103A (en) | Thermometry and relevant device | |
| CN107045623A (en) | A kind of method of the indoor dangerous situation alarm based on human body attitude trace analysis | |
| CN107657244B (en) | Human body falling behavior detection system based on multiple cameras and detection method thereof | |
| CN110956118B (en) | Target object detection method and device, storage medium and electronic device | |
| JP2004054376A (en) | Method and device for estimating group attribute | |
| CN104284084B (en) | Image processing equipment, image processing method and computer storage medium | |
| CN110321767A (en) | Image acquiring apparatus and method, behavior analysis system and storage medium | |
| WO2017098265A1 (en) | Method and apparatus for monitoring | |
| CN108875614A (en) | It is a kind of that detection method is fallen down based on deep learning image procossing | |
| Goudelis et al. | Fall detection using history triple features | |
| Pramerdorfer et al. | Fall detection based on depth-data in practice | |
| Pinitkan et al. | Abnormal activity detection and notification platform for real-time ad hoc network | |
| WO2020217812A1 (en) | Image processing device that recognizes state of subject and method for same | |
| CN115695734A (en) | Infrared thermal imaging protection monitoring method, device, equipment, system and medium | |
| CN108549838B (en) | Auxiliary supervision method based on visual system | |
| Liu et al. | MonkeyTrail: A scalable video-based method for tracking macaque movement trajectory in daily living cages | |
| Khan et al. | Video analytic for fall detection from shape features and motion gradients | |
| CN112836549A (en) | User information detection method and system and electronic equipment | |
| CN115331312A (en) | Human body state detection method, device and equipment and readable storage medium |
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: 20200204 |
|
| RJ01 | Rejection of invention patent application after publication |