CN114694073B - Intelligent detection method, device, storage medium and equipment for wearing condition of safety belt - Google Patents
Intelligent detection method, device, storage medium and equipment for wearing condition of safety belt Download PDFInfo
- Publication number
- CN114694073B CN114694073B CN202210355160.2A CN202210355160A CN114694073B CN 114694073 B CN114694073 B CN 114694073B CN 202210355160 A CN202210355160 A CN 202210355160A CN 114694073 B CN114694073 B CN 114694073B
- Authority
- CN
- China
- Prior art keywords
- safety belt
- person
- detection
- data
- video data
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/251—Fusion techniques of input or preprocessed data
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
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)
- Alarm Systems (AREA)
Abstract
The embodiment of the invention provides an intelligent detection method, device and equipment for the wearing condition of a safety belt and a storage medium, wherein the method comprises the following steps: acquiring video data of a scene; transmitting the video data to a data fusion processing center to obtain optimized data; and carrying out feature extraction on the optimized data by combining a target detection network and a key point detection network to obtain a detection result of the wearing condition of the safety belt. Through the scheme, a user can acquire operation field picture video data by using an unmanned aerial vehicle or a monitoring camera and the like, the video data are sent to a data fusion processing center to obtain optimized data, and then feature extraction is carried out on the optimized data by combining a target detection network and a key point detection network to obtain a detection result of the wearing condition of the safety belt, so that intelligent detection of the wearing condition of the safety belt is realized, the detection efficiency and accuracy of the wearing condition of the safety belt are improved, and the manpower resource waste and the safety risk of personnel in field supervision are reduced.
Description
Technical Field
The embodiment of the invention relates to the technical field of intelligent detection of power construction sites, in particular to an intelligent detection method, an intelligent detection device, intelligent detection equipment and intelligent detection storage media for wearing conditions of safety belts.
Background
Along with the rapid development of power system construction, the safe operation level of a power grid needs to be ensured and the service level of a customer needs to be improved, but the intelligent development of the current power system is still in a starting stage, and the construction site often needs to be manually subjected to safety supervision. The high-altitude operation safety belt is also called a whole-body safety belt or a five-point type safety belt, and the novel national standard GB6095-2009 prescribes that the material is processed by using terylene and higher-strength woven belts. The whole-body safety belt is an important protective article for preventing falling casualties of high-rise operators, the wearing of the safety belt is manually monitored for guaranteeing the life and property safety of the operators, but the traditional manual monitoring for the wearing condition of the safety belt is low in efficiency, the condition that the monitoring is not in place is easy to generate, hidden danger is bred, and even safety accidents are caused, so that at present, an intelligent detection and identification method for the wearing condition of the safety belt of the operators on an electric power construction site is not available, and the operators cannot timely and accurately know whether the wearing condition of the safety belt of the operators is illegal or not and remind the operators.
Disclosure of Invention
Aiming at the problems, the embodiment of the invention provides an intelligent detection method, device, equipment and storage medium for the wearing condition of the safety belt, so that the intelligent detection of the wearing condition of the safety belt on the electric power construction site is realized, the manpower resource waste and the safety risk of the site supervision are greatly reduced, and the supervision efficiency and effect are improved.
In a first aspect, an intelligent detection method for a wearing condition of a safety belt according to an embodiment of the present invention is characterized by including:
acquiring video data of a scene;
transmitting the video data to a data fusion processing center to obtain optimized data;
and carrying out feature extraction on the optimized data by combining a target detection network and a key point detection network to obtain a detection result of the wearing condition of the safety belt.
Preferably, the video data of the field picture is acquired through unmanned aerial vehicle equipment and/or a monitoring camera.
Preferably, the sending the video data to a data fusion processing center to obtain optimized data includes:
and carrying out optimized screening and re-processing on the video data.
The feature extraction of the optimized data by combining the target detection network and the key point detection network to obtain a detection result of the wearing condition of the safety belt comprises the following steps:
acquiring characteristic information;
the characteristic information is fused through the target detection network, a detection area of a person is obtained, and hanging points, hanging rings, a vertical main belt and vertical auxiliary belts in the safety belt are detected within the detection area of the person;
positioning skeleton point information of the person in the detection area of the person through the key point detection network, and calculating the real position orientation of the person according to the skeleton point information;
and obtaining a detection result of the wearing condition of the safety belt according to the skeleton point information and the orientation.
Further, the acquiring the feature information includes:
intercepting a single frame picture of the optimized data;
performing size normalization on the single-frame picture, performing slicing operation on the single-frame picture to obtain a stacked picture, and performing convolution operation on the stacked picture to obtain a sampling feature picture;
and carrying out feature extraction on the sampling feature map to obtain the feature information.
Further, the locating the skeleton point information of the person in the detection area of the person, calculating the true position orientation of the person according to the skeleton point information, includes:
the orientation is calculated from the pixel length between the shoulders and buttocks in the detection area of the person.
Further, the detecting result of the wearing condition of the safety belt according to the skeleton point information and the orientation includes:
judging whether the hanging points and hanging rings of the safety belt exceed a preset threshold according to the relation between the skeleton point information and the orientation, and if yes, outputting a result of illegal wearing of the safety belt.
In a second aspect, an embodiment of the present invention further provides an intelligent detection device for a wearing condition of a safety belt, including:
the video data acquisition module is used for acquiring video data of the field picture;
the data processing module is used for sending the video data to a data fusion processing center so as to obtain optimized data;
the detection result acquisition module is used for carrying out feature extraction on the optimized data by combining the target detection network and the key point detection network so as to obtain a detection result of the wearing condition of the safety belt.
In a third aspect, an embodiment of the present invention further provides a storage medium, which is a computer readable storage medium, where a computer program is stored, where the computer program is executed to implement the steps of the method for detecting an entry situation of a person in a foundation area as described above.
In a fourth aspect, the present invention also provides a computer device, which includes a processor, a memory, and a computer program stored in the memory and executable on the processor, where the processor executes the computer program to implement the steps of the method for detecting an entry condition of a person in a foundation area as described above.
According to the embodiment of the invention, through the scheme, a user can acquire the video data of the operation scene picture by using an unmanned aerial vehicle or a monitoring camera and the like, the video data is sent to the data fusion processing center to obtain the optimized data, and then the optimized data is subjected to characteristic extraction by combining the target detection network and the key point detection network to obtain the detection result of the wearing condition of the safety belt, so that the intelligent detection of the wearing condition of the safety belt is realized, the detection efficiency and accuracy of the wearing condition of the safety belt are improved, and the manpower resource waste and the safety risk of personnel in the scene supervision are reduced.
Drawings
FIG. 1 is a schematic flow chart of an intelligent detection method for the wearing condition of a safety belt according to an embodiment of the present invention;
FIG. 2 is a flow chart of a method for intelligently detecting the wearing condition of a safety belt according to an embodiment of the present invention;
FIG. 3 is a flow chart of a method for intelligently detecting the wearing condition of a safety belt according to an embodiment of the present invention;
FIG. 4 is a schematic structural view of an intelligent detection device for detecting the wearing condition of a safety belt according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a storage medium according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a computer device according to an embodiment of the present invention.
Detailed Description
The invention is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting thereof. It should be further noted that, for convenience of description, only some, but not all of the structures related to the present invention are shown in the drawings.
Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although a flowchart depicts steps as a sequential process, many of the steps may be implemented in parallel, concurrently, or with other steps. Furthermore, the order of the steps may be rearranged. The process may be terminated when its operations are completed, but may have additional steps not included in the figures. The processes may correspond to methods, functions, procedures, subroutines, and the like.
Furthermore, the terms "first," "second," and the like, may be used herein to describe various directions, acts, steps, or elements, etc., but these directions, acts, steps, or elements are not limited by these terms. These terms are only used to distinguish one direction, action, step or element from another direction, action, step or element. For example, a first processing unit may be referred to as a second processing unit, and similarly, a second processing unit may be referred to as a first processing unit, without departing from the scope of the present application. Both the first processing unit and the second processing unit are processing units, but they are not the same processing unit. The terms "first," "second," and the like, are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present invention, the meaning of "plurality" means at least two, for example, two, three, etc., unless specifically defined otherwise.
Fig. 1 is a flow chart of an intelligent detection method for a belt wearing condition, which is provided by an embodiment of the present invention, and is applicable to a scenario of intelligent detection of a belt wearing condition of a person on a construction site.
As shown in fig. 1, the intelligent detection method for the wearing condition of the safety belt provided by the embodiment of the invention includes:
s10, acquiring video data of a scene;
s20, sending the video data to a data fusion processing center to obtain optimized data;
and S30, combining the target detection network and the key point detection network to perform feature extraction on the optimized data so as to obtain a detection result of the wearing condition of the safety belt.
In the embodiment of the invention, the video data of the field picture can be acquired through unmanned aerial vehicle equipment and/or a monitoring camera. The unmanned aerial vehicle equipment is preferably a four-rotor professional unmanned aerial vehicle DJIM300RTK or a eidolon 4/4pro Dajiang unmanned aerial vehicle and the like, the four-rotor professional unmanned aerial vehicle can be provided with a visible light load holder camera, the load adopts a Buddhist X5S holder camera, the unmanned aerial vehicle has the advantages of compact design, flexible expansion, remarkable optimization of an intelligent control system and flight performance, functions of flight and data safety and the like are added, and the unmanned aerial vehicle has good performance in the process of collecting conditions of a construction site. The intelligent unmanned aerial vehicle with the intelligent smart 4 is excellent in flexibility and portability, the camera of the Phantom 4Pro is provided with a 1 inch 2000 ten thousand pixel image sensor, and is matched with a rearview vision sensor and infrared sensors on two sides of a body, so that 5-direction environment recognition and 4-direction obstacle avoidance capability are achieved, the safety is higher, the flying is more intelligent, and a user can flexibly select unmanned aerial vehicle equipment to acquire video data of a field picture. Preferably, the video data includes real-time video data and offline video data. The embodiment of the invention can also respectively carry out the subsequent steps for the real-time video data and the offline video data.
Further, sending the video data to a data fusion processing center to obtain optimized data includes: and after the video data are sent to a data fusion processing center, optimizing and screening the video data. By optimizing the weight of the video data, only one of the same video data is reserved as the optimized data, repeated operations caused by repeated video data can be reduced, and the processing efficiency is improved. And then, combining the target detection network and the key point detection network to perform feature extraction on the optimized data so as to obtain a detection result of the wearing condition of the safety belt.
Further, referring to fig. 2, in the embodiment of the present invention, the step S30 of performing feature extraction on the optimized data by combining the target detection network and the key point detection network to obtain a detection result of the wearing condition of the safety belt includes:
s31, acquiring characteristic information;
s32, fusing the characteristic information through the target detection network to obtain a detection area of a person, and detecting hanging points, hanging rings, vertical main belts and vertical auxiliary belts in the safety belt within the detection area of the person;
s33, positioning skeleton point information of the person in the detection area of the person through the key point detection network, and calculating the real position orientation of the person according to the skeleton point information;
s34, detecting results of the wearing condition of the safety belt are obtained according to the skeleton point information, the orientation and the hanging points, hanging rings, vertical main belts and vertical auxiliary positions in the safety belt.
In the embodiment of the invention, the characteristic information is firstly obtained from the video data.
Referring to fig. 3, the step S31 of acquiring feature information includes:
s31a, intercepting a single-frame picture of the optimized data;
s31b, performing size normalization on the single-frame picture, performing slicing operation on the single-frame picture to obtain a stacked picture, and performing convolution operation on the stacked picture to obtain a sampling feature picture;
and S31c, carrying out feature extraction on the sampling feature map to obtain the feature information.
Firstly, intercepting a single frame picture of the optimized data, namely intercepting a single frame picture in the video subjected to optimized screening treatment as a single frame picture, secondly, carrying out size normalization on the single frame picture, and carrying out slicing operation on the single frame picture to obtain a stacked picture. Specifically, a value can be taken from every other pixel in the single frame picture, the picture is sliced, and four pictures are obtained, so W, H information is concentrated into a channel space, an input channel is expanded by 4 times, namely, the spliced picture becomes 12 channels relative to the original RGB three-channel mode, the speed of a network is improved, no information loss is caused, then the four pictures are stacked to obtain the stacked picture, and then the stacked picture is subjected to convolution operation to obtain a sampling feature map.
In the embodiment of the invention, the stacked pictures are convolved to obtain the sampling feature map, convolution and repeated residual error operations are carried out on the stacked pictures to obtain the base layer, then the CSP module in yolov5 is adopted to divide the special mapping of the base layer into two parts, and then the two parts are combined through a cross-stage hierarchical structure, so that the calculated amount is reduced, and meanwhile, the accuracy can be ensured. Then, extracting the features of the sampled feature map to obtain the feature information may include: and respectively carrying out normalization processing on the characteristic quantities output by the network, so that the data distribution of each characteristic is converted into a mean value 0 and a variance 1. The range of the eigenvalue numbers is controlled between 0 and 1. The sensitivity to some super parameters is reduced, the generalization capability is improved, and better convergence speed and convergence effect can be obtained. Pooling the feature graphs with any size to obtain a fixed number of features, combining the features obtained by pooling to obtain the feature number with a fixed length, continuously downsampling the feature points, providing a stack of feature layers with high semantic content, and then upsampling again to enlarge the length and width of the feature layers again, and stacking the feature layers with the same length and width as those in the downsampling.
After the characteristic information is acquired, the characteristic information is fused through the target detection network, so that a detection area of a person is acquired, and hanging points, hanging rings, vertical main belts and vertical auxiliary belts in the safety belt are detected in the detection area of the person. The characteristic information is fused to extract the characteristics of the person so as to obtain the detection area range of the person, and whether hanging points and hanging rings in the safety belt exist or not is searched in the detection area of the person; further, in the detection area of the person, U-net may be employed as a semantic division model to detect the vertical main band and the vertical incidental band. All the numerical values of all the image pixels are normalized, the prediction probability that each pixel in the image is a vertical main band or a vertical auxiliary band is obtained through a U-net semantic segmentation model, at the moment, all the values are multiplied by 255 to convert the vertical main band and the vertical auxiliary band probability map into a preliminary rope segmentation gray map, then a maximum inter-class variance algorithm is adopted to conduct fine threshold segmentation, the accuracy of rope edge segmentation is improved, and pixels with smaller gray values are removed, so that the vertical main band and the vertical auxiliary band can be obtained.
And then, positioning skeleton point information of the person in the detection area of the person obtained in the previous step through the key point detection network, and calculating the real position orientation of the person according to the skeleton point information. Wherein, first locate the skeleton point information of the person in the detection area of the person, the said skeleton point information can include head, shoulder, buttock and knee joint, etc.. In practical implementation, by comparing the variances of the position coordinates of the head-shoulder, shoulder-hip and hip-knee joints in all persons using a probability statistical method, the variance of the shoulder-hip is the smallest, which means that calculating the position orientation of the person in the real world is the most accurate according to the pixel length relationship of the shoulder-hip, for example, assuming that all persons are regarded as the same height or the same body size, using a skeleton to extract pixel information of the shoulder-hip on an image, and thus the true position orientation of the person can be obtained. Specifically, in the image, a triangle relationship can be formed between the shoulders and the buttocks, the orientation of the person can be obtained through the pixel length relationship of the triangle, the distance from the person to the camera can be calculated according to the principle of similar triangles, the position information of the person in the image is mapped onto a three-dimensional coordinate graph, and the specific position of the person is located.
And finally, obtaining a detection result of the wearing condition of the safety belt according to the skeleton point information, the orientation and the hanging points, hanging rings, the vertical main belt and the vertical attached positions in the safety belt. In the embodiment of the present invention, the step S34 of obtaining the detection result of the wearing condition of the safety belt according to the skeleton point information, the orientation, and the hanging points, the hanging rings, the vertical main belt, and the vertical auxiliary positions in the safety belt includes: when the hanging point and the hanging ring of the safety belt exceed a preset threshold, outputting a result of illegal wearing of the safety belt; and outputting a result that the safety belt is suspected to be unworn when the vertical main belt or the vertical auxiliary belt is detected. The vertical main belt and the vertical accessory are the main belt and the accessory of the safety belt which are not wound on the tower from the body, the body and the tower are connected, the main belt and the accessory are only hung on the body, and the tower is completely free of connecting points, namely is regarded as being vertical on the body, and is the vertical main belt and the vertical accessory at the moment, so that when the vertical main belt or the vertical accessory is detected, the condition that the safety belt is suspected to be unworn can be considered, and the result that the safety belt is suspected to be unworn can be output.
In addition, in the embodiment of the invention, an alarm signal can be generated after a result of illegal wearing of the safety belt is output or a result that the safety belt is suspected to be unworn is output, the alarm signal is uploaded to the interaction center, a detection report can be generated according to the detection result, and the detection report is uploaded to the interaction center. In actual implementation, the interaction center may include: a user interface, a data management module, and a report display module. The data management module can display and manage video data on a user interface, the report display module can be used for displaying alarm information or detection reports generated according to the alarm signals, and further, a user can download the detection reports and can be used for carrying out subsequent behavior evaluation or summarization of related personnel. In addition, the interaction center in the embodiment of the invention further comprises an alarm module, wherein the alarm module can be used for generating an alarm prompt according to the alarm signal and timely notifying a manager or a mobile device worn by a field person.
According to the embodiment of the invention, through the scheme, a user can acquire the video data of the operation scene picture by using an unmanned aerial vehicle or a monitoring camera and the like, the video data is sent to the data fusion processing center to obtain the optimized data, and then the optimized data is subjected to characteristic extraction by combining the target detection network and the key point detection network to obtain the detection result of the wearing condition of the safety belt, so that the intelligent detection of the wearing condition of the safety belt is realized, the detection efficiency and accuracy of the wearing condition of the safety belt are improved, and the manpower resource waste and the safety risk of personnel in the scene supervision are reduced.
Fig. 4 is a schematic structural diagram of an intelligent detection device for a wearing condition of a safety belt according to an embodiment of the present invention, and referring to fig. 4, the intelligent detection device for a wearing condition of a safety belt according to an embodiment of the present invention includes: a video data acquisition module 1, a data processing module 2 and a detection result acquisition module 3.
The video data acquisition module 1 is used for acquiring video data of a field picture;
the data processing module 2 is used for sending the video data to a data fusion processing center so as to obtain optimized data;
the detection result obtaining module 3 is configured to perform feature extraction on the optimized data in combination with a target detection network and a key point detection network to obtain a detection result of the wearing condition of the safety belt.
Preferably, the video data acquisition module 1-bit unmanned aerial vehicle device and/or the monitoring camera can acquire video data of the field picture through the unmanned aerial vehicle device and/or the monitoring camera.
Further, the data processing module is specifically configured to perform optimized screening processing on the video data after sending the video data to a data fusion processing center. By optimizing the weight of the video data, only one of the same video data is reserved as the optimized data, repeated operations caused by repeated video data can be reduced, and the processing efficiency is improved. And then, combining the target detection network and the key point detection network to perform feature extraction on the optimized data so as to obtain a detection result of the wearing condition of the safety belt.
In one embodiment, the detection result acquisition module 3 includes:
the first processing unit is used for acquiring the characteristic information;
the second processing unit is used for fusing the characteristic information through the target detection network to obtain a detection area of a person and detecting hanging points, hanging rings, vertical main belts and vertical auxiliary belts in the safety belt in the detection area of the person;
the third processing unit is used for positioning skeleton point information of the person in the detection area of the person through the key point detection network and calculating the real position orientation of the person according to the skeleton point information;
and the fourth processing unit is used for obtaining the detection result of the wearing condition of the safety belt according to the skeleton point information, the orientation and the hanging points, hanging rings, vertical main belts and vertical attached positions in the safety belt.
Specifically, the first processing unit further includes:
the first processing subunit is used for intercepting single-frame pictures of the optimized data;
the second processing subunit is used for carrying out size normalization on the single-frame picture, carrying out slicing operation on the single-frame picture to obtain a stacked picture, and carrying out convolution operation on the stacked picture to obtain a sampling feature picture;
and the third processing subunit is used for carrying out feature extraction on the sampling feature map to obtain the feature information.
In particular, the third processing unit is in particular adapted to calculate the orientation from a pixel length between the shoulders and the buttocks in the detection area of the person.
The fourth processing unit is specifically configured to: when the hanging point and the hanging ring of the safety belt exceed a preset threshold, outputting a result of illegal wearing of the safety belt; and outputting a result that the safety belt is suspected to be unworn when the vertical main belt or the vertical auxiliary belt is detected.
Firstly, intercepting a single frame picture of the optimized data, namely intercepting a single frame picture in the video subjected to optimized screening treatment as a single frame picture, secondly, carrying out size normalization on the single frame picture, and carrying out slicing operation on the single frame picture to obtain a stacked picture. Specifically, a value can be taken from every other pixel in the single frame picture, the picture is sliced, and four pictures are obtained, so W, H information is concentrated into a channel space, an input channel is expanded by 4 times, namely, the spliced picture becomes 12 channels relative to the original RGB three-channel mode, the speed of a network is improved, no information loss is caused, then the four pictures are stacked to obtain the stacked picture, and then the stacked picture is subjected to convolution operation to obtain a sampling feature map.
In the embodiment of the invention, the stacked pictures are convolved to obtain the sampling feature map, convolution and repeated residual error operations are carried out on the stacked pictures to obtain the base layer, then the CSP module in yolov5 is adopted to divide the special mapping of the base layer into two parts, and then the two parts are combined through a cross-stage hierarchical structure, so that the calculated amount is reduced, and meanwhile, the accuracy can be ensured. Then, extracting the features of the sampled feature map to obtain the feature information may include: and respectively carrying out normalization processing on the characteristic quantities output by the network, so that the data distribution of each characteristic is converted into a mean value 0 and a variance 1. The range of the eigenvalue numbers is controlled between 0 and 1. The sensitivity to some super parameters is reduced, the generalization capability is improved, and better convergence speed and convergence effect can be obtained. Pooling the feature graphs with any size to obtain a fixed number of features, combining the features obtained by pooling to obtain the feature number with a fixed length, continuously downsampling the feature points, providing a stack of feature layers with high semantic content, and then upsampling again to enlarge the length and width of the feature layers again, and stacking the feature layers with the same length and width as those in the downsampling.
After the characteristic information is acquired, the characteristic information is fused through the target detection network, so that a detection area of a person is acquired, and hanging points, hanging rings, vertical main belts and vertical auxiliary belts in the safety belt are detected in the detection area of the person. The characteristic information is fused to extract the characteristics of the person so as to obtain the detection area range of the person, and whether hanging points and hanging rings in the safety belt exist or not is searched in the detection area of the person; further, in the detection area of the person, U-net may be employed as a semantic division model to detect the vertical main band and the vertical incidental band. All the numerical values of all the image pixels are normalized, the prediction probability that each pixel in the image is a vertical main band or a vertical auxiliary band is obtained through a U-net semantic segmentation model, at the moment, all the values are multiplied by 255 to convert the vertical main band and the vertical auxiliary band probability map into a preliminary rope segmentation gray map, then a maximum inter-class variance algorithm is adopted to conduct fine threshold segmentation, the accuracy of rope edge segmentation is improved, and pixels with smaller gray values are removed, so that the vertical main band and the vertical auxiliary band can be obtained.
And then, positioning skeleton point information of the person in the detection area of the person obtained in the previous step through the key point detection network, and calculating the real position orientation of the person according to the skeleton point information. Wherein, first locate the skeleton point information of the person in the detection area of the person, the said skeleton point information can include head, shoulder, buttock and knee joint, etc.. In practical implementation, by comparing the variances of the position coordinates of the head-shoulder, shoulder-hip and hip-knee joints in all persons using a probability statistical method, the variance of the shoulder-hip is the smallest, which means that calculating the position orientation of the person in the real world is the most accurate according to the pixel length relationship of the shoulder-hip, for example, assuming that all persons are regarded as the same height or the same body size, using a skeleton to extract pixel information of the shoulder-hip on an image, and thus the true position orientation of the person can be obtained. Specifically, in the image, a triangle relationship can be formed between the shoulders and the buttocks, the orientation of the person can be obtained through the pixel length relationship of the triangle, the distance from the person to the camera can be calculated according to the principle of similar triangles, the position information of the person in the image is mapped onto a three-dimensional coordinate graph, and the specific position of the person is located.
And finally, obtaining a detection result of the wearing condition of the safety belt according to the skeleton point information, the orientation and the hanging points, hanging rings, the vertical main belt and the vertical attached positions in the safety belt. Outputting a result of illegal wearing of the safety belt when the hanging point and the hanging ring of the safety belt exceed a preset threshold value; and outputting a result that the safety belt is suspected to be unworn when the vertical main belt or the vertical auxiliary belt is detected. The vertical main belt and the vertical accessory are the main belt and the accessory of the safety belt which are not wound on the tower from the body, the body and the tower are connected, the main belt and the accessory are only hung on the body, and the tower is completely free of connecting points, namely is regarded as being vertical on the body, and is the vertical main belt and the vertical accessory at the moment, so that when the vertical main belt or the vertical accessory is detected, the condition that the safety belt is suspected to be unworn can be considered, and the result that the safety belt is suspected to be unworn can be output.
In addition, in the embodiment of the present invention, a fifth processing unit may be further included, which is specifically configured to generate an alarm signal after outputting a result of illegal wearing of the safety belt or outputting a result that the safety belt is suspected to be unworn, and upload the alarm signal to the interaction center, or generate a detection report according to the detection result, and upload the detection report to the interaction center. In actual implementation, the interaction center may include: a user interface, a data management module, and a report display module. The data management module can display and manage video data on a user interface, the report display module can be used for displaying alarm information or detection reports generated according to the alarm signals, and further, a user can download the detection reports and can be used for carrying out subsequent behavior evaluation or summarization of related personnel. In addition, the interaction center in the embodiment of the invention further comprises an alarm module, wherein the alarm module can be used for generating an alarm prompt according to the alarm signal and timely notifying a manager or a mobile device worn by a field person.
Referring to fig. 5, an embodiment of the present invention further provides a storage medium 100, in which a computer program 200 is stored, which when run on a computer, causes the computer to perform the method for intelligent detection of the wearing condition of the seat belt described in the above embodiment.
Referring to fig. 6, an embodiment of the present invention further provides a device 300 including instructions, which is a computer device, when running on the device 300, to cause the device 300 to perform, through a processor 400 disposed therein, the intelligent detection method of the belt wearing condition described in the above embodiment.
It will be appreciated by those skilled in the art that the method, apparatus and device for intelligent detection of belt wear as described herein and the apparatus referred to above for performing one or more of the methods described herein. These devices may be specially designed and constructed for the required purposes, or may comprise known devices in general purpose computers. These devices have computer programs or applications stored therein that are selectively activated or reconfigured. Such a computer program may be stored in a device (e.g., a computer) readable medium or any type of medium suitable for storing electronic instructions and respectively coupled to a bus, including, but not limited to, any type of disk (including floppy disks, hard disks, optical disks, CD-ROMs, and magneto-optical disks), ROMs (Read-Only memories), RAMs (Random Access Memory, random access memories), EPROMs (Erasable Programmable Read-Only memories), EEPROMs (Electrically Erasable Programmable Read-Only memories), flash memories, magnetic cards, or optical cards. That is, a readable medium includes any medium that stores or transmits information in a form readable by a device (e.g., a computer).
Note that the above is only a preferred embodiment of the present invention and the technical principle applied. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, while the invention has been described in connection with the above embodiments, the invention is not limited to the embodiments, but may be embodied in many other equivalent forms without departing from the spirit or scope of the invention, which is set forth in the following claims.
Claims (7)
1. The intelligent detection method for the wearing condition of the safety belt is characterized by comprising the following steps of:
acquiring video data of a scene;
transmitting the video data to a data fusion processing center to obtain optimized data;
performing feature extraction on the optimized data by combining a target detection network and a key point detection network to obtain a detection result of the wearing condition of the safety belt;
the feature extraction of the optimized data by combining the target detection network and the key point detection network to obtain a detection result of the wearing condition of the safety belt comprises the following steps:
acquiring characteristic information;
the characteristic information is fused through the target detection network, a detection area of a person is obtained, and hanging points, hanging rings, a vertical main belt and vertical auxiliary belts in the safety belt are detected in the detection area of the person;
positioning skeleton point information of the person in the detection area of the person through the key point detection network, and calculating the real position orientation of the person according to the skeleton point information;
obtaining a detection result of the wearing condition of the safety belt according to the skeleton point information, the orientation and the hanging points, hanging rings, vertical main belts and vertical auxiliary positions in the safety belt;
positioning skeleton point information of a person in a detection area of the person through the key point detection network, and calculating the true position orientation of the person according to the skeleton point information, wherein the method comprises the following steps:
calculating the orientation from the pixel length between the shoulders and buttocks in the detection area of the person;
the method for detecting the wearing condition of the safety belt according to the skeleton point information, the orientation and the hanging points, hanging rings, vertical main belts and vertical attached positions in the safety belt comprises the following steps:
when the hanging point and the hanging ring of the safety belt exceed a preset threshold, outputting a result of illegal wearing of the safety belt;
and outputting a result that the safety belt is suspected to be unworn when the vertical main belt or the vertical auxiliary belt is detected.
2. The method according to claim 1, wherein the video data of the live view is acquired by a drone device and/or a surveillance camera.
3. The method of claim 1, wherein said sending the video data to a data fusion processing center to obtain optimized data comprises:
and after the video data are sent to a data fusion processing center, optimizing and screening the video data.
4. The method of claim 1, wherein the obtaining feature information comprises:
intercepting a single frame picture of the optimized data;
performing size normalization on the single-frame picture, performing slicing operation on the single-frame picture to obtain a stacked picture, and performing convolution operation on the stacked picture to obtain a sampling feature picture;
and carrying out feature extraction on the sampling feature map to obtain the feature information.
5. Intelligent detection device of condition is worn to safety belt, its characterized in that includes:
the video data acquisition module is used for acquiring video data of the field picture;
the data processing module is used for sending the video data to a data fusion processing center so as to obtain optimized data;
the detection result acquisition module is used for carrying out feature extraction on the optimized data by combining a target detection network and a key point detection network so as to obtain a detection result of the wearing condition of the safety belt;
the detection result acquisition module comprises:
the first processing unit is used for acquiring the characteristic information;
the second processing unit is used for fusing the characteristic information through the target detection network to obtain a detection area of a person and detecting hanging points, hanging rings, vertical main belts and vertical auxiliary belts in the safety belt in the detection area of the person;
the third processing unit is used for positioning skeleton point information of the person in the detection area of the person through the key point detection network and calculating the real position orientation of the person according to the skeleton point information;
the fourth processing unit is used for obtaining a detection result of the wearing condition of the safety belt according to the skeleton point information, the orientation and the hanging points, hanging rings, vertical main belts and vertical attached positions in the safety belt;
the third processing unit is specifically configured to: calculating the orientation from the pixel length between the shoulders and buttocks in the detection area of the person;
the fourth processing unit is specifically configured to: when the hanging point and the hanging ring of the safety belt exceed a preset threshold, outputting a result of illegal wearing of the safety belt; and outputting a result that the safety belt is suspected to be unworn when the vertical main belt or the vertical auxiliary belt is detected.
6. A storage medium having a computer program stored thereon, characterized in that it is a computer-readable storage medium having a computer program stored thereon, which computer program, when executed, realizes the steps of the method for intelligent detection of a belt wearing situation according to any one of claims 1 to 4.
7. An apparatus comprising a processor, a memory and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the method for intelligently detecting the wearing condition of a seat belt as claimed in any one of claims 1 to 4 when the computer program is executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210355160.2A CN114694073B (en) | 2022-04-06 | 2022-04-06 | Intelligent detection method, device, storage medium and equipment for wearing condition of safety belt |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210355160.2A CN114694073B (en) | 2022-04-06 | 2022-04-06 | Intelligent detection method, device, storage medium and equipment for wearing condition of safety belt |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114694073A CN114694073A (en) | 2022-07-01 |
| CN114694073B true CN114694073B (en) | 2023-06-06 |
Family
ID=82143307
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210355160.2A Active CN114694073B (en) | 2022-04-06 | 2022-04-06 | Intelligent detection method, device, storage medium and equipment for wearing condition of safety belt |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114694073B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115049678A (en) * | 2022-08-17 | 2022-09-13 | 南昌工程学院 | Transmission line corona discharge image segmentation method based on night ultraviolet imaging technology |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111222478A (en) * | 2020-01-11 | 2020-06-02 | 深圳奥腾光通***有限公司 | Construction site safety protection detection method and system |
| CN112215138A (en) * | 2020-10-12 | 2021-01-12 | 中国石油大学(华东) | Deep learning-based violation detection method for low hanging height of safety belt |
| CN113971829A (en) * | 2021-10-28 | 2022-01-25 | 广东律诚工程咨询有限公司 | Intelligent detection method, device, equipment and storage medium for wearing condition of safety helmet |
| CN114120205A (en) * | 2021-12-02 | 2022-03-01 | 云南电网有限责任公司信息中心 | Target detection and image recognition method for safety belt fastening of distribution network operators |
| CN114187542A (en) * | 2021-11-29 | 2022-03-15 | 国网福建省电力有限公司建设分公司 | Insulating glove detection method and system in electric power scene |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008002838A (en) * | 2006-06-20 | 2008-01-10 | Takata Corp | System for detecting vehicle occupant, actuator control system, and vehicle |
| CN113870205A (en) * | 2021-09-18 | 2021-12-31 | 中国联合网络通信集团有限公司 | Safety belt wearing detection method and device, electronic equipment and storage medium |
| CN113807331A (en) * | 2021-11-19 | 2021-12-17 | 中化学交通建设集团有限公司 | Safety belt wearing identification method and device based on image identification technology |
| CN114187543A (en) * | 2021-11-29 | 2022-03-15 | 国网福建省电力有限公司建设分公司 | Safety belt detection method and system in high-altitude power operation scene |
-
2022
- 2022-04-06 CN CN202210355160.2A patent/CN114694073B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111222478A (en) * | 2020-01-11 | 2020-06-02 | 深圳奥腾光通***有限公司 | Construction site safety protection detection method and system |
| CN112215138A (en) * | 2020-10-12 | 2021-01-12 | 中国石油大学(华东) | Deep learning-based violation detection method for low hanging height of safety belt |
| CN113971829A (en) * | 2021-10-28 | 2022-01-25 | 广东律诚工程咨询有限公司 | Intelligent detection method, device, equipment and storage medium for wearing condition of safety helmet |
| CN114187542A (en) * | 2021-11-29 | 2022-03-15 | 国网福建省电力有限公司建设分公司 | Insulating glove detection method and system in electric power scene |
| CN114120205A (en) * | 2021-12-02 | 2022-03-01 | 云南电网有限责任公司信息中心 | Target detection and image recognition method for safety belt fastening of distribution network operators |
Non-Patent Citations (1)
| Title |
|---|
| 基于LS-DYNA的汽车安全带固定点强度分析;廉振红;《内燃机与配件》;第39-41页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114694073A (en) | 2022-07-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101715001B1 (en) | Display system for safety evaluation in construction sites using of wearable device, and thereof method | |
| CN106341661B (en) | Patrol robot | |
| CN106851229B (en) | Security and protection intelligent decision method and system based on image recognition | |
| KR101988356B1 (en) | Smart field management system through 3d digitization of construction site and analysis of virtual construction image | |
| CN111753780B (en) | Transformer substation violation detection system and violation detection method | |
| CN114694073B (en) | Intelligent detection method, device, storage medium and equipment for wearing condition of safety belt | |
| CN112183472A (en) | Method for detecting whether test field personnel wear work clothes or not based on improved RetinaNet | |
| CN112184628B (en) | Infrared duplex wave image and cloud early warning system and method for flood prevention and danger detection of dike | |
| CN112784821A (en) | Building site behavior safety detection and identification method and system based on YOLOv5 | |
| CN114665608B (en) | Intelligent sensing inspection system and method for transformer substation | |
| KR102388454B1 (en) | Pine wilt disease Spreading Prevention System and method using Semantic Segmentation | |
| CN106960027B (en) | The UAV Video big data multidate association analysis method of spatial information auxiliary | |
| CN115861915A (en) | Fire fighting access monitoring method, fire fighting access monitoring device and storage medium | |
| CN113505704B (en) | Personnel safety detection method, system, equipment and storage medium for image recognition | |
| CN113628172A (en) | Intelligent detection algorithm for personnel handheld weapons and smart city security system | |
| CN117372875A (en) | Aerial remote sensing target identification method | |
| CN115953815A (en) | Monitoring method and device for infrastructure site | |
| CN114708544A (en) | Intelligent violation monitoring helmet based on edge calculation and monitoring method thereof | |
| CN111985331B (en) | Detection method and device for preventing trade secret from being stolen | |
| WO2022173809A1 (en) | Computer vision systems and methods for detecting power line hazards from imagery | |
| CN114912536A (en) | Target identification method based on radar and double photoelectricity | |
| CN113989886A (en) | Crew identity verification method based on face recognition | |
| CN116912721B (en) | Power distribution network equipment body identification method and system based on monocular stereoscopic vision | |
| CN117152689B (en) | River channel target detection method and system based on vision | |
| Kang et al. | Development of Smart Helmet for Monitoring Construction Resources Based on Image Matching Method. |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |