CN116862244B - Industrial field vision AI analysis and safety pre-warning system and method - Google Patents
Industrial field vision AI analysis and safety pre-warning system and method Download PDFInfo
- Publication number
- CN116862244B CN116862244B CN202311128741.3A CN202311128741A CN116862244B CN 116862244 B CN116862244 B CN 116862244B CN 202311128741 A CN202311128741 A CN 202311128741A CN 116862244 B CN116862244 B CN 116862244B
- Authority
- CN
- China
- Prior art keywords
- safety
- model
- accident
- analysis
- area
- 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
- 238000004458 analytical method Methods 0.000 title claims abstract description 57
- 238000000034 method Methods 0.000 title claims abstract description 32
- 230000004438 eyesight Effects 0.000 title claims abstract description 17
- 238000012502 risk assessment Methods 0.000 claims abstract description 17
- 230000000391 smoking effect Effects 0.000 claims abstract description 12
- 231100000817 safety factor Toxicity 0.000 claims abstract description 7
- 230000008447 perception Effects 0.000 claims abstract description 5
- 238000013527 convolutional neural network Methods 0.000 claims description 44
- 238000007726 management method Methods 0.000 claims description 37
- 238000001514 detection method Methods 0.000 claims description 34
- 238000012549 training Methods 0.000 claims description 27
- 230000008859 change Effects 0.000 claims description 13
- 230000006378 damage Effects 0.000 claims description 13
- 238000012805 post-processing Methods 0.000 claims description 12
- 230000009471 action Effects 0.000 claims description 10
- 230000006399 behavior Effects 0.000 claims description 10
- 238000013480 data collection Methods 0.000 claims description 7
- 230000006872 improvement Effects 0.000 claims description 7
- 231100000572 poisoning Toxicity 0.000 claims description 7
- 230000000607 poisoning effect Effects 0.000 claims description 7
- 208000027418 Wounds and injury Diseases 0.000 claims description 6
- 208000014674 injury Diseases 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000004140 cleaning Methods 0.000 claims description 5
- 230000005764 inhibitory process Effects 0.000 claims description 5
- 238000005065 mining Methods 0.000 claims description 5
- 238000013450 outlier detection Methods 0.000 claims description 5
- 230000000007 visual effect Effects 0.000 claims 3
- 238000012544 monitoring process Methods 0.000 description 8
- 230000008901 benefit Effects 0.000 description 5
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000036544 posture Effects 0.000 description 2
- 230000001629 suppression Effects 0.000 description 2
- 206010000369 Accident Diseases 0.000 description 1
- 238000012300 Sequence Analysis Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000004297 night vision Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- 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
- 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
-
- 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
- G06V40/20—Movements or behaviour, e.g. gesture recognition
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B31/00—Predictive alarm systems characterised by extrapolation or other computation using updated historic 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
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Human Resources & Organizations (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Emergency Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Human Computer Interaction (AREA)
- Strategic Management (AREA)
- Operations Research (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Development Economics (AREA)
- Alarm Systems (AREA)
Abstract
The invention discloses an industrial field vision AI analysis and safety pre-warning system and method, comprising a system platform layer, a network layer and a perception layer, wherein the system platform layer comprises: the AI algorithm platform is used for carrying out algorithm analysis on videos, pictures and personnel trajectories which are accessed into the factory and affect the safety factors, and the algorithm analysis comprises the steps of safety helmet wearing, tumbling, work clothes wearing, smoking, calling and hand lifting for help seeking; the safety helmet management platform is used for storing and managing the data of voice call, video call and personnel positioning of the safety helmet. The safety risk assessment is carried out through the mathematical expected value, the frequency of occurrence of the accident in a certain area can be predicted by calculating the probability of occurrence of the specific accident in each area, a certain reference basis can be given to a manager, attention is paid to the accident frequently occurring in the area, the accident cause is easier to find, and therefore the field facilities are continuously perfected.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an industrial field vision AI analysis and safety early warning system and method.
Background
In the production of large industrial factories, in order to ensure the normal operation of enterprises and realize economic benefit targets, the safety of the production process is very important. The operation processes such as a plurality of operation processes, material handling and equipment facilities have higher dangers, so that the potential safety hazard is more, the phenomenon of wearing no safety helmet, working clothes and the like is frequent, and the pollution is higher; in addition, the probability of occurrence of the safety accidents in each area is not systematically analyzed, so that the safety accidents can be repeatedly generated in the same area, the advance prediction and the precaution measures cannot be made, the continuous occurrence of the accidents are caused, the risk is difficult to reduce, the economic benefit and the reputation of a factory are greatly influenced, and the use of AI analysis to promote the safety and environmental protection management is a necessary requirement for the modern development of industrial enterprises.
Disclosure of Invention
In order to solve the technical problems in the background art, the invention provides an industrial field vision AI analysis and safety early warning system and method capable of analyzing and detecting safety-affecting behaviors through an AI intelligent algorithm and evaluating and analyzing safety risks. The specific technical scheme is as follows:
in a first aspect, the present invention provides an industrial field vision AI analysis and security pre-warning system, including a system platform layer, a network layer, and a perception layer, wherein the system platform layer includes:
the AI algorithm platform is used for carrying out algorithm analysis on videos, pictures and personnel trajectories which are accessed into the factory and affect the safety factors, and the algorithm analysis comprises the steps of safety helmet wearing, tumbling, work clothes wearing, smoking, calling and hand lifting for help seeking;
the safety helmet management platform is used for storing and managing the data of voice call, video call and personnel positioning of the safety helmet;
the video management platform is used for storing and managing voice call, video call and picture data of the camera;
the safety management and control platform is used for approval of special dangerous operation, approval of operation ticket, safety risk assessment analysis, performance management work injury management, standing account management and operation planning; the method is also used for counting the security events and identifying time periods or areas with abnormally high numbers of the security events;
the sensing layer comprises a sensor group, wherein the sensor group comprises a camera group;
the network layer includes a 5G network.
Preferably, the detection of the helmet wearing and the work wear identification by adopting a convolutional neural network and a YOLO algorithm comprises the following steps:
the method for identifying the wearing of the safety helmet and the wearing of the working clothes adopts convolutional neural network and YOLO algorithm detection and comprises the following steps:
training a CNN model by using marked target object image data, and training a CNN model to accurately judge whether a target object exists in an image or not and the position of the target object;
training a YOLO model, namely training a YOLO model by using the marked target object image data and the target object position information to realize target detection and predict the position and the type of the target object;
integrating the trained CNN model and the YOLO model, wherein the CNN model is used for extracting image features, inputting the image features into the YOLO model for target detection and classification, and outputting the position information of a target object by the YOLO model;
post-processing, namely, executing a non-maximum value inhibition post-processing step on a plurality of object detection results output by the YOLO so as to screen out a final target object detection result;
the security management platform is further configured to perform security risk assessment analysis, the security risk assessment analysis including:
data collection and classification: counting the number of safety accidents in each area within a period of time, wherein the safety accidents comprise actions of falling, lifting hands for help, falling aloft, fire disaster, poisoning and artificial destruction, and classifying;
dividing the area: dividing the factory floor into a plurality of areas;
counting the accident number of each area: counting the number of each area according to the type of the security accident;
calculating an expected value: the expected value of each accident type for each region is calculated, and for a particular region i and a particular accident type j, the probability of occurrence of that accident type for that region is first calculated, i.e., P (accident j│ Region(s) i ) Calculating expected values of each accident type, and using the following formula expected value E(accident) j│ Region(s) i ) =p (accident) j│ Region(s) i ) x total number of incidents; wherein the total number of incidents is the total number of incidents that occur in all areas.
Calculating the probability of occurrence of specific accidents in each area, calculating the corresponding expected value of the specific safety accidents occurring or expected to occur in the current or future time period, analyzing and evaluating the expected value of each area, and establishing a safety analysis relevance prediction model; pre-collecting security event data including different types of security incidents, date and time information occurring in each region; analyzing the collected data to know occurrence frequencies of different areas and different types of accidents, and performing pretreatment steps of data cleaning, outlier detection and missing data filling; then, establishing a relevance prediction model, and selecting proper characteristics to describe the conditions of each region and each security event type, wherein the conditions comprise the characteristics, time characteristics and event types of the regions; an association rule mining algorithm Apriori algorithm is used to find associations between different events.
Preferably, the probability of occurrence of a specific accident in each area is calculated, the corresponding expected value of the specific safety accident occurring or expected to occur in the current or future time period is calculated, the expected value of each area is analyzed and evaluated, and the establishment of a safety analysis relevance prediction model is made.
Preferably, new data is continuously input into the model according to the established prediction model, calculation or prediction of event results is performed through the input model, and the model is repeatedly adjusted.
Preferably, if the change rate of the prediction model in the preset time period does not reach the minimum value, the system sends out early warning information to remind that the safety accident problem is improved and does not reach the standard.
In a second aspect, the present invention also provides an industrial field vision AI analysis and safety precaution method, including the steps of:
carrying out algorithm analysis on videos, pictures and personnel trajectories which are accessed into a factory and affect safety factors, wherein the algorithm analysis comprises the steps of detecting behaviors of wearing safety helmets, falling, wearing work clothes, smoking, calling, lifting hands and asking for help;
storing and managing the voice call, the video call and the personnel positioning data of the safety helmet;
storing and managing voice call, video call and picture data of the camera;
performing approval of special dangerous operation, approval of operation ticket, security risk assessment analysis, performance management work injury management, account standing management and operation planning; counting the security events and identifying a time period or area in which the number of the security events is abnormally high;
the method for identifying the wearing of the safety helmet and the wearing of the working clothes adopts convolutional neural network and YOLO algorithm detection and comprises the following steps:
training a CNN model, namely training a CNN model by using marked safety helmet image data, so that whether safety helmet exists in the image or not and the position of the safety helmet can be accurately judged;
training a YOLO model, namely training a YOLO model by using marked safety helmet image data and safety helmet position information to realize target detection and predict the position and the type of the safety helmet;
integrating the trained CNN model and the YOLO model, wherein the CNN model is used for extracting image features, inputting the image features into the YOLO model for target detection and classification, and outputting the position information of the safety helmet by the YOLO model;
post-processing, namely, executing a non-maximum value inhibition post-processing step on a plurality of object detection results output by the YOLO so as to screen out a final safety helmet detection result;
the security risk assessment analysis includes:
data collection and classification: counting the number of safety accidents in each area within a period of time, wherein the safety accidents comprise actions of falling, lifting hands for help, falling aloft, fire disaster, poisoning and artificial destruction, and classifying;
dividing the area: dividing the factory floor into a plurality of areas;
counting the accident number of each area: counting the number of each area according to the type of the security accident;
calculating an expected value: the expected value of each accident type for each region is calculated, and for a particular region i and a particular accident type j, the probability of occurrence of that accident type for that region is first calculated, i.e., P (accident j│ Region(s) i ) The expected value of each accident type was calculated, and the following formula expected value E (accident j│ Region(s) i ) =p (accident) j│ Region(s) i ) x total number of incidents; wherein the total number of incidents is the total number of incidents that occur in all areas.
Preferably, the method comprises the steps of:
and calculating the probability of occurrence of specific accidents in each area, calculating the corresponding expected value of the specific safety accidents occurring or expected to occur in the current or future time period, analyzing and evaluating the expected value of each area, and establishing a safety analysis relevance prediction model. Pre-collecting security event data including different types of security incidents, date and time information occurring in each region; analyzing the collected data to know occurrence frequencies of different areas and different types of accidents, and performing pretreatment steps of data cleaning, outlier detection and missing data filling; then, establishing a relevance prediction model, and selecting proper characteristics to describe the conditions of each region and each security event type, wherein the conditions comprise the characteristics, time characteristics and event types of the regions; an association rule mining algorithm Apriori algorithm is used to find associations between different events.
Preferably, the method comprises the steps of:
continuously inputting new data into the model according to the established prediction model, calculating or predicting event results through the input model, and repeatedly adjusting the model; if the change rate of the prediction model in the preset time period does not reach the minimum value, the system sends out early warning information to remind that the safety accident problem improvement does not reach the standard.
The beneficial effects of the invention are as follows:
through the action of utilizing AI algorithm discernment helmet wearing, falling, worker's clothes wearing, smoking, calling, lifting hand and asking for help in the factory, realize intelligent accurate discernment, reduce through the neglect omission that manual monitoring leads to, can effectively reduce the human cost, improve economic benefits.
The field video information and the call voice information are acquired by using various cameras, such as a control ball, so that the real-time situation of the field is collected in time, and real-time tracking and tracking are facilitated.
The convolutional neural network and the YOLO algorithm are adopted for detection on the wearing of the safety helmet and the wearing of the work clothes, so that the existence or the absence of the safety helmet and the work clothes can be detected with high accuracy, false alarms and missing alarms can be reduced, and the safety is improved. Convolutional neural networks and YOLO algorithms work well under a variety of environmental conditions, including illumination changes, background clutter, and the like. The convolutional neural network and the YOLO algorithm are applied to safety helmet wearing and work wear identification, so that safety, automation and efficiency of an industrial field can be improved, and potential dangerous situations are reduced. Is very beneficial for ensuring that the staff obeys the safety regulations and reducing the accident risk.
The method has the advantages that the mathematical expected value is used for carrying out security risk assessment, the probability of occurrence of a specific accident in each area is calculated, the corresponding expected value of the specific accident occurring or expected to occur in the current or future time period is calculated, the expected value of each area is analyzed and assessed, the number of times of occurrence of the accident in a certain area can be predicted, a manager can be given a certain reference basis, frequent accidents in the area are focused, accident reasons are easier to find, field facilities are continuously perfected, and a good modification iteration loop is formed.
Drawings
FIG. 1 is a schematic block diagram of an industrial field vision AI analysis and safety warning system of the invention.
FIG. 2 is a flow chart of an industrial field vision AI analysis and safety precaution method of the invention.
FIG. 3 is another flow chart of an industrial field vision AI analysis and safety warning method of the invention.
FIG. 4 is a flow chart of an industrial field vision AI analysis and safety warning method of the invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. On the contrary, the invention is intended to cover any alternatives, modifications, equivalents, and variations as may be included within the spirit and scope of the invention as defined by the appended claims. Further, in the following detailed description of the present invention, certain specific details are set forth in order to provide a better understanding of the present invention. The present invention will be fully understood by those skilled in the art without the details described herein.
Referring to fig. 1 to 4, in one aspect, the present invention provides an industrial field vision AI analysis and security pre-warning system, which includes a system platform layer 10, a network layer 20, and a perception layer 30, wherein the system platform layer 10 includes: the AI algorithm platform 12 is used for performing algorithm analysis on videos, pictures and personnel trajectories which are accessed into the factory and affect the safety factors, and detecting behaviors of helmet wearing, falling, worker wearing, smoking, calling and hand lifting for help; the helmet management platform 14 is used for storing and managing the data of voice call, video call and personnel positioning of the helmet; a video management platform 13 for storing and managing voice call, video call, and picture data of the camera; the safety management and control platform 11 is used for approval of special dangerous operation, approval of operation ticket, safety risk assessment analysis, performance management work injury management, account management and operation planning; the method is also used for counting the security events and identifying time periods or areas with abnormally high numbers of the security events; the sensing layer 30 comprises a set of sensors, wherein the set of sensors comprises a set of cameras; the network layer 20 comprises a 5G network.
In the embodiment, the actions of wearing the safety helmet, falling, wearing the work clothes, smoking, making a call and asking for help are detected through various recognition algorithms, wherein the recognition of the safety helmet and the work clothes can be performed through a convolutional neural network and a YOLO algorithm; the tumbling detection can use a behavior recognition and target detection algorithm, and can be used for recognizing the human body gesture in real time by monitoring key points of the human body or using a deep learning model, such as a model based on CNN and a circulating neural network, so as to judge whether the tumbling occurs. The behavior recognition of smoking and making a call can use a behavior recognition method in combination with monitoring hand gestures and object recognition. By monitoring the position and posture of the hand, and whether smoking articles or mobile phones exist in the image, whether a person is smoking or making a call is judged. Gesture estimation and behavior recognition methods can be used for the recognition of the help-seeking behavior. The key points of the human body, particularly the positions and the postures of the hands are monitored to judge whether the hand lifting and help seeking actions occur. The algorithm identification is carried out by collecting the picture and video information through the camera group of the system, and the picture and video information is sent to the AI algorithm platform for analysis.
The safety helmet management platform 14 is connected with the safety helmet through a 5G network, receives positioning information, video information, voice information and the like sent by the safety helmet, and displays the personnel track with the 5G intelligent safety helmet in real time through an electronic map. Specifically, the helmet may use a 5G smart helmet MT5000.
The video management platform 13 is used for receiving the picture information and video information of various cameras such as a control ball, a full-color night vision camera, a high-temperature furnace camera and the like, and storing and managing the voice call, the video call and the picture data of the camera.
The safety management and control platform 11 is used for approval of special dangerous operation and approval of operation ticket, and comprises an approval operation type, an application unit, an applicant, operation content, an operation certificate number (unique), a certificate name, an operation safety education department, an operation safety education person, an approval list number (unique), a finished check-up person, an operation mode, a shift, an operation responsible person, an operation supervisor, an operation level, information related to special dangerous operation, a risk identification result, an operation person and the like, wherein the operation worker needs to submit a dangerous operation approval application or an operation ticket approval application on a mobile phone terminal in advance, and the operation responsible person can give approval on the safety management platform. And when the operation is completed, the task state is fed back to the operation task of the safety control platform.
In the system platform, a perception layer 30 is in information interaction with a system platform layer 10 through a network layer 20, and then the system platform layer 10 is connected with an application layer, wherein the application layer is a terminal device such as a monitoring large screen, a mobile phone computer and the like.
Further, the detection of the helmet wear and the work wear identification by adopting a convolutional neural network and a YOLO algorithm comprises the following steps:
training a convolutional neural network CNN model, namely training a CNN model by using marked target object image data, so that whether a target object exists in an image or not and the position of the target object can be accurately judged;
training a YOLO model, namely training a YOLO model by using the marked target object image data and the target object position information to realize target detection and predict the position and the type of the target object;
integrating the trained CNN model and the YOLO model, wherein the CNN model is used for extracting image features, inputting the image features into the YOLO model for target detection and classification, and outputting the position information of a target object by the YOLO model;
and (3) post-processing, namely, performing a non-maximum value suppression post-processing step on a plurality of object detection results output by the YOLO so as to screen out a final target object detection result.
In the present embodiment, specific recognition is
Step 1: data preparation: image datasets containing helmet and non-helmet scenes are collected and annotated. The callout should include the location of the helmet (bounding box) and a category label in each image. This data will be used to train and validate the model.
Step 2: training convolutional neural networks: using the collected data set, a convolutional neural network is trained so that it can detect the helmets in the image. Some common CNN architecture may be used, such as res net, VGG, etc. During the training process, the network will learn to extract features from the images and locate the helmet.
Step 3: integrating YOLO algorithm: after the convolutional neural network training is completed, integrating the YOLO algorithm. The YOLO algorithm may help detect multiple objects and generate bounding boxes and class probabilities for each object.
Step 4: fine tuning of the model: after the trained convolutional neural network is combined with the YOLO algorithm, the whole model can be finely adjusted as required, including the adjustment of the super parameters of the model, such as confidence threshold, non-maximum suppression threshold and the like.
Step 5: testing and verifying: the overall model was tested and validated using a separate test dataset to evaluate its performance. Ensuring that the model is able to accurately detect the presence of the helmet and locate its position. By combining the CNN and the YOLO algorithm, a real-time safety helmet recognition system can be realized, and whether safety helmets and work clothes in an industrial field are worn or not can be rapidly and accurately detected and positioned. The combination of CNN and YOLO algorithms for monitoring helmets and workshops can achieve high efficiency, multi-target detection capability, high precision and stronger adaptability.
Further, the security management platform is further configured to perform security risk assessment analysis, where the security risk assessment analysis includes:
data collection and classification: counting the number of safety accidents in each area within a period of time, wherein the safety accidents comprise actions of falling, lifting hands for help, falling aloft, fire disaster, poisoning and artificial destruction, and classifying;
dividing the area: dividing the factory floor into a plurality of areas;
counting the accident number of each area: counting the number of each area according to the type of the security accident;
calculating an expected value: the expected value of each accident type for each region is calculated, and for a particular region i and a particular accident type j, the probability of occurrence of that accident type for that region is first calculated, i.e., P (accident j│ Region(s) i ) The expected value of each accident type was calculated, and the following formula expected value E (accident j│ Region(s) i ) =p (accident) j│ Region(s) i ) x total number of incidents; wherein the total number of incidents is the total number of incidents that occur in all areas.
In the present embodiment of the present invention, in the present embodiment,
step 1: data collection and classification: safety accident data is collected in relation to each zone. Each accident is classified into different types such as fall, hand lifting, help seeking, high altitude falling, fire, poisoning, manual damage and the like. The data should include the area of the incident occurrence, the date and the type of incident.
Step 2: dividing the area: the division of the industrial plant into a plurality of zones ensures that the boundaries of each zone are clearly identifiable, which can be done according to the actual industrial field layout and zone division.
Step 3: counting the accident number of each area: for each zone, the number of different types of security incidents is calculated, for example:
area a fall incident: 10 get up
Area a lifts hands for help: 5 take up
Area B high altitude fall: 3 get up
Area B fire: 2 get up
Step 4: calculating an expected value: calculating the expected value of each accident type in each area, wherein the calculation of the expected value can be realized by the following steps: for a particular region i and a particular accident type j, the occurrence (probability) of that accident type for that region is first calculated, i.e., P (accident j│ Region(s) i ). Can be obtained by counting the number of incidents of this type in this area divided by the total number of incidents. The expected value for each accident type is calculated using the following formula:
expected value E (accident) j│ Region(s) i ) =p (accident) j│ Region(s) i ) x total number of incidents.
For example, for region a and the expected value calculation of the fall incident:
p (fall |zone a) = (assuming a total of 15 accidents
E ({ fall } | { region a }) =10/15×15=10
The above formula can calculate the expected value of each accident type in each area, so that the manager can know the average occurrence frequency of each accident type in each area, and is helpful to take appropriate safety measures and precautions to reduce the possibility of subsequent continuous occurrence; for example, the area A frequently falls down, the floor friction of the area A can be enhanced, or improvement measures such as anti-skid shoes can be suggested for workers.
Further, S24, by calculating the probability of occurrence of a specific accident in each area, calculating the expected value corresponding to the specific accident occurring or expected to occur in the current or future time period, analyzing and evaluating the expected value of each area, and establishing a security analysis relevance prediction model.
In this embodiment, a large amount of security event data including different types of security incidents, date and time information, and the like occurring in each area are collected in advance. The data are historical data. Analyzing the collected data to know the occurrence frequency of different areas and different types of accidents, and performing preprocessing steps such as data cleaning, outlier detection, missing data filling and the like to ensure the quality and the integrity of the data. And further builds a relevance prediction model, and selects proper characteristics (or attributes) to describe each area and the condition of each security event type, including the characteristics of the area (such as the number of workers and the number of devices), the time characteristics (such as date and time period), the event type and the like. The goal of feature engineering is to extract features that have the highest relevance to security events. Association rule mining algorithms (such as Apriori algorithms) are used to find associations between different events. For example, a common pattern of some specific event combinations may be viewed to predict the correlation between them; for example, a fall accident often occurs in the area a, and a fire accident often occurs in the area B. In addition, time sequence analysis can be used for analyzing time sequence data, so that the change trend of different event types along with time is explored, and management staff can be helped to know whether certain periodic relevance exists or not; for example, area C is subject to theft incidents during certain periods of time, according to which the manager can strengthen the defender and caretaker. After the model is built, it needs to be evaluated and optimized. The historical data is used to test the performance of the model and compare the predicted results of the model with what actually happens to determine the accuracy and reliability of the model. Parameters or characteristics of the model can be adjusted according to actual needs to improve the performance of the model. Once a reliable relevance prediction model is established, it can be applied to the prediction of security events in the current or future time period. The model may identify potential security risks and take action in advance to reduce the likelihood of an accident.
Establishing the relevance prediction model has the following beneficial effects:
(1) early warning in advance: the model can identify potential safety problems in advance so as to take preventive measures in time and reduce accident risks.
(2) And (3) resource optimization: by predicting the occurrence of an accident, management personnel can more effectively allocate resources, and ensure that security measures are fully utilized.
(3) Data driven decision: the model makes decisions based on data and facts, influences of subjective judgment are reduced, and scientificity of the decisions is improved.
(4) Continuous improvement: by constantly monitoring the performance and results of the model, continuous improvements and optimizations can be made to adapt to constantly changing safety environments.
Further, S25, new data is continuously input into the model according to the established prediction model, calculation or prediction of event results is performed through the input model, and the model is repeatedly adjusted.
In this embodiment, the built prediction model continuously inputs new data into the model, so as to input the result output after the current optimization environment into the model, so as to facilitate observation of the change trend of the event after the rectification of the rectification measure, and whether the event tends to be subtracted or not can be observed, and if the event tends to be subtracted does not exist, the rectification strategy can be adjusted in time, so as to reduce the possibility of reoccurrence.
Further, S26, if the rate of change of the prediction model in the preset time period does not reach the minimum value, the system sends out early warning information to remind that the safety accident problem improvement does not reach the standard.
In this embodiment, a period of time may be preset in the system, and the rate of change is calculated by comparing model performance indicators between two time points at the same time:
(1) selecting performance indexes: an index that evaluates the performance of the model is determined, and for a relevance prediction model, the index may include accuracy, recall, F1 score, and the like.
(2) Two time points were selected: two time points for comparison are selected. Typically these points in time should be relatively close in order to monitor changes in model performance.
(3) Calculating performance indexes: the values of the selected performance indicators are calculated at two time points, respectively. For example, the model may be calculated to an accuracy of 0.85 at time point 1 and 0.80 at time point 2.
(4) Calculating the change rate: rate of change= (performance index 2-performance index 1)/performance index 1.
(5) Setting a threshold value: a threshold is determined that represents a minimum acceptable level of performance variation. The threshold value is selected according to the actual requirement. For example, the threshold may be set to-0.05, indicating that if the model performance drops by 5% or more, the system will send out warning information.
(6) Monitoring and triggering early warning: the rate of change of the model performance is calculated periodically and compared to a set threshold. If the change rate is lower than the threshold value, the performance of the model is reduced to an unacceptable level, and the system should trigger early warning information to remind the safety accident problem.
On the other hand, based on the same conception, the invention also provides an industrial field vision AI analysis and safety pre-warning method, wherein the working principle of the industrial field vision AI analysis and safety pre-warning system provided by the invention is the same as or similar to that of the industrial field vision AI analysis and safety pre-warning system, and therefore, repeated parts are not repeated. The method comprises the following steps:
carrying out algorithm analysis on videos, pictures and personnel trajectories which are accessed into a factory and affect safety factors, wherein the algorithm analysis comprises the steps of detecting behaviors of wearing safety helmets, falling, wearing work clothes, smoking, calling, lifting hands and asking for help;
storing and managing the voice call, the video call and the personnel positioning data of the safety helmet;
storing and managing voice call, video call and picture data of the camera;
performing approval of special dangerous operation, approval of operation ticket, security risk assessment analysis, performance management work injury management, account standing management and operation planning; statistics of the security events and identification of time periods or areas in which the number of the security events is abnormally high;
s10, detecting the wearing of the safety helmet and the wearing identification of the working clothes by adopting a convolutional neural network and a YOLO algorithm, wherein the method comprises the following steps:
training a CNN model, namely training a CNN model by using marked safety helmet image data, so that whether safety helmet exists in the image or not and the position of the safety helmet can be accurately judged;
s11, training a YOLO model, namely training a YOLO model by using marked safety helmet image data and safety helmet position information to realize target detection and predict the position and the type of the safety helmet;
s12, integrating a trained CNN model and a YOLO model, wherein the CNN model is used for extracting image features, inputting the image features into the YOLO model for target detection and classification, and outputting the position information of the safety helmet by the YOLO model;
s13, post-processing, namely, performing a non-maximum value inhibition post-processing step on a plurality of object detection results output by the YOLO so as to screen out a final safety helmet detection result.
Further, the method comprises the steps of: s20, security risk assessment analysis comprises:
data collection and classification: counting the number of safety accidents in each area within a period of time, wherein the safety accidents comprise actions of falling, lifting hands for help, falling aloft, fire disaster, poisoning and artificial destruction, and classifying;
s21, area division: dividing the factory floor into a plurality of areas;
s22, counting the accident number of each area: counting the number of each area according to the type of the security accident;
s23, calculating an expected value: the expected value of each accident type for each region is calculated, and for a particular region i and a particular accident type j, the probability of occurrence of that accident type for that region is first calculated, i.e., P (accident j│ Region(s) i ) The expected value of each accident type was calculated, and the following formula expected value E (accidentTherefore, it is j│ Region(s) i ) =p (accident) j│ Region(s) i ) x total number of incidents; wherein the total number of incidents is the total number of incidents that occur in all areas.
Further, the method comprises the steps of:
s24, calculating the probability of occurrence of specific accidents in each area, calculating the corresponding expected value of the specific safety accidents occurring or expected to occur in the current or future time period, analyzing and evaluating the expected value of each area, and establishing a safety analysis relevance prediction model.
Further, the method comprises the steps of:
s25, continuously inputting new data into the model according to the established prediction model, calculating or predicting event results through the input model, and repeatedly adjusting the model;
s26, if the change rate of the prediction model in the preset time period does not reach the minimum value, the system sends out early warning information to remind that the safety accident problem is not reduced yet.
The above embodiments are merely illustrative of the principles of the present invention and its effectiveness, and are not intended to limit the invention. Modifications and variations may be made to the above-described embodiments by those skilled in the art without departing from the spirit and scope of the invention. Accordingly, it is intended that all equivalent modifications and variations of the invention be covered by the claims, which are within the ordinary skill of the art, be within the spirit and scope of the present disclosure.
Claims (5)
1. The industrial field vision AI analysis and safety pre-warning system is characterized by comprising a system platform layer, a network layer and a perception layer, wherein the system platform layer comprises:
the AI algorithm platform is used for carrying out algorithm analysis on videos, pictures and personnel trajectories which are accessed into the factory and affect the safety factors, and the algorithm analysis comprises the steps of safety helmet wearing, tumbling, work clothes wearing, smoking, calling and hand lifting for help seeking;
the safety helmet management platform is used for storing and managing the data of voice call, video call and personnel positioning of the safety helmet;
the video management platform is used for storing and managing voice call, video call and picture data of the camera;
the safety management and control platform is used for approval of special dangerous operation, approval of operation ticket, safety risk assessment analysis, performance management, work injury management, standing account management and operation planning; the method is also used for counting the security events and identifying time periods or areas with abnormally high numbers of the security events;
the sensing layer comprises a sensor group, wherein the sensor group comprises a camera group;
the network layer comprises a 5G network;
the method for identifying the wearing of the safety helmet and the wearing of the working clothes adopts convolutional neural network and YOLO algorithm detection and comprises the following steps:
training a CNN model by using marked target object image data, and training a CNN model to accurately judge whether a target object exists in an image or not and the position of the target object;
training a YOLO model, namely training a YOLO model by using the marked target object image data and the target object position information to realize target detection and predict the position and the type of the target object;
integrating the trained CNN model and the YOLO model, wherein the CNN model is used for extracting image features, inputting the image features into the YOLO model for target detection and classification, and outputting the position information of a target object by the YOLO model;
post-processing, namely, executing a non-maximum value inhibition post-processing step on a plurality of object detection results output by the YOLO so as to screen out a final target object detection result;
the security management platform is further configured to perform security risk assessment analysis, the security risk assessment analysis including:
data collection and classification: counting the number of safety accidents in each area within a period of time, wherein the safety accidents comprise actions of falling, lifting hands for help, falling aloft, fire disaster, poisoning and artificial destruction, and classifying;
dividing the area: dividing the factory floor into a plurality of areas;
counting the accident number of each area: counting the number of each area according to the type of the security accident;
calculating an expected value: the expected value of each accident type for each region is calculated, and for a particular region i and a particular accident type j, the probability of occurrence of that accident type for that region is first calculated, i.e., P (accident j│ Region(s) i ) The expected value of each accident type was calculated, and the following formula expected value E (accident j│ Region(s) i ) =p (accident) j│ Region(s) i ) x total number of incidents; wherein the total number of incidents is the total number of incidents that occur in all areas;
calculating the probability of occurrence of specific accidents in each area, calculating the corresponding expected value of the specific safety accidents occurring or expected to occur in the current or future time period, analyzing and evaluating the expected value of each area, and establishing a safety analysis relevance prediction model; pre-collecting security event data including different types of security incidents, date and time information occurring in each region; analyzing the collected data to know occurrence frequencies of different areas and different types of accidents, and performing pretreatment steps of data cleaning, outlier detection and missing data filling; then, establishing a relevance prediction model, and selecting proper characteristics to describe the conditions of each region and each security event type, wherein the conditions comprise the characteristics, time characteristics and event types of the regions; an association rule mining algorithm Apriori algorithm is used to find associations between different events.
2. The industrial field visual AI analysis and safety precaution system of claim 1, wherein new data is continually input into the model based on the established predictive model, calculation or prediction of event results is performed by the input model, and the model is iteratively adjusted.
3. The industrial field visual AI analysis and safety precaution system of claim 2, wherein if the rate of change of the predictive model within the predetermined time period does not reach a minimum value, the system sends a precaution message to alert the safety accident that the improvement does not reach standard.
4. An industrial field vision AI analysis and safety pre-warning method is characterized by comprising the following steps:
carrying out algorithm analysis on videos, pictures and personnel trajectories which are accessed into a factory and affect safety factors, wherein the algorithm analysis comprises the steps of detecting behaviors of wearing safety helmets, falling, wearing work clothes, smoking, calling, lifting hands and asking for help;
storing and managing the voice call, the video call and the personnel positioning data of the safety helmet;
storing and managing voice call, video call and picture data of the camera;
performing approval of special dangerous operation, approval of operation ticket, security risk assessment analysis, performance management work injury management, account standing management and operation planning; counting the security events and identifying a time period or area in which the number of the security events is abnormally high;
the method for identifying the wearing of the safety helmet and the wearing of the working clothes adopts convolutional neural network and YOLO algorithm detection and comprises the following steps:
training a CNN model by using marked target object image data, and training a CNN model to accurately judge whether a target object exists in an image or not and the position of the target object;
training a YOLO model, namely training a YOLO model by using the marked target object image data and the target object position information to realize target detection and predict the position and the type of the target object;
integrating the trained CNN model and the YOLO model, wherein the CNN model is used for extracting image features, inputting the image features into the YOLO model for target detection and classification, and outputting the position information of a target object by the YOLO model;
post-processing, namely, executing a non-maximum value inhibition post-processing step on a plurality of object detection results output by the YOLO so as to screen out a final target object detection result;
the security risk assessment analysis includes:
data collection and classification: counting the number of safety accidents in each area within a period of time, wherein the safety accidents comprise actions of falling, lifting hands for help, falling aloft, fire disaster, poisoning and artificial destruction, and classifying;
dividing the area: dividing the factory floor into a plurality of areas;
counting the accident number of each area: counting the number of each area according to the type of the security accident;
calculating an expected value: the expected value of each accident type for each region is calculated, and for a particular region i and a particular accident type j, the probability of occurrence of that accident type for that region is first calculated, i.e., P (accident j│ Region(s) i ) The expected value of each accident type was calculated, and the following formula expected value E (accident j│ Region(s) i ) =p (accident) j│ Region(s) i ) x total number of incidents; wherein the total number of incidents is the total number of incidents that occur in all areas;
calculating the probability of occurrence of specific accidents in each area, calculating the corresponding expected value of the specific safety accidents occurring or expected to occur in the current or future time period, analyzing and evaluating the expected value of each area, and establishing a safety analysis relevance prediction model; pre-collecting security event data including different types of security incidents, date and time information occurring in each region; analyzing the collected data to know occurrence frequencies of different areas and different types of accidents, and performing pretreatment steps of data cleaning, outlier detection and missing data filling; then, establishing a relevance prediction model, and selecting proper characteristics to describe the conditions of each region and each security event type, wherein the conditions comprise the characteristics, time characteristics and event types of the regions; an association rule mining algorithm Apriori algorithm is used to find associations between different events.
5. The industrial field visual AI analysis and safety precaution method of claim 4, comprising the steps of:
continuously inputting new data into the model according to the established prediction model, calculating or predicting event results through the input model, and repeatedly adjusting the model; if the change rate of the prediction model in the preset time period does not reach the minimum value, the system sends out early warning information to remind that the safety accident problem improvement does not reach the standard.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311128741.3A CN116862244B (en) | 2023-09-04 | 2023-09-04 | Industrial field vision AI analysis and safety pre-warning system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311128741.3A CN116862244B (en) | 2023-09-04 | 2023-09-04 | Industrial field vision AI analysis and safety pre-warning system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116862244A CN116862244A (en) | 2023-10-10 |
| CN116862244B true CN116862244B (en) | 2024-03-22 |
Family
ID=88221983
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311128741.3A Active CN116862244B (en) | 2023-09-04 | 2023-09-04 | Industrial field vision AI analysis and safety pre-warning system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116862244B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4850973B1 (en) * | 2010-09-14 | 2012-01-11 | ギアヌーヴ株式会社 | Construction site labor management system and server |
| CN104901838A (en) * | 2015-06-23 | 2015-09-09 | 中国电建集团成都勘测设计研究院有限公司 | Enterprise network safety event management system and method thereof |
| CN112183979A (en) * | 2020-09-18 | 2021-01-05 | 浙江省安全生产科学研究院 | Hazardous chemical substance loading and unloading safety risk assessment method based on multi-source information fusion |
| CN112434827A (en) * | 2020-11-23 | 2021-03-02 | 南京富岛软件有限公司 | Safety protection identification unit in 5T fortune dimension |
| CN113971666A (en) * | 2021-10-29 | 2022-01-25 | 贵州电网有限责任公司 | Power transmission line machine inspection image self-adaptive identification method based on depth target detection |
| CN115001940A (en) * | 2022-05-27 | 2022-09-02 | 北京双湃智安科技有限公司 | Association security situation analysis method based on artificial intelligence |
| CN115169777A (en) * | 2022-05-09 | 2022-10-11 | 国家电投集团内蒙古能源有限公司 | Intelligent supervision system for safety operation |
| CN115660297A (en) * | 2021-07-07 | 2023-01-31 | 北京宜通科创科技发展有限责任公司 | Automatic AI early warning system and method for construction site safety |
| WO2023116507A1 (en) * | 2021-12-22 | 2023-06-29 | 北京沃东天骏信息技术有限公司 | Target detection model training method and apparatus, and target detection method and apparatus |
-
2023
- 2023-09-04 CN CN202311128741.3A patent/CN116862244B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4850973B1 (en) * | 2010-09-14 | 2012-01-11 | ギアヌーヴ株式会社 | Construction site labor management system and server |
| CN104901838A (en) * | 2015-06-23 | 2015-09-09 | 中国电建集团成都勘测设计研究院有限公司 | Enterprise network safety event management system and method thereof |
| CN112183979A (en) * | 2020-09-18 | 2021-01-05 | 浙江省安全生产科学研究院 | Hazardous chemical substance loading and unloading safety risk assessment method based on multi-source information fusion |
| CN112434827A (en) * | 2020-11-23 | 2021-03-02 | 南京富岛软件有限公司 | Safety protection identification unit in 5T fortune dimension |
| CN115660297A (en) * | 2021-07-07 | 2023-01-31 | 北京宜通科创科技发展有限责任公司 | Automatic AI early warning system and method for construction site safety |
| CN113971666A (en) * | 2021-10-29 | 2022-01-25 | 贵州电网有限责任公司 | Power transmission line machine inspection image self-adaptive identification method based on depth target detection |
| WO2023116507A1 (en) * | 2021-12-22 | 2023-06-29 | 北京沃东天骏信息技术有限公司 | Target detection model training method and apparatus, and target detection method and apparatus |
| CN115169777A (en) * | 2022-05-09 | 2022-10-11 | 国家电投集团内蒙古能源有限公司 | Intelligent supervision system for safety operation |
| CN115001940A (en) * | 2022-05-27 | 2022-09-02 | 北京双湃智安科技有限公司 | Association security situation analysis method based on artificial intelligence |
Non-Patent Citations (2)
| Title |
|---|
| 本杰明•普朗什（Benjamin Planche）.《计算机视觉实战 基于TensorFlow 2》.机械工业出版社,2021,110-111. * |
| 陈艳艳.《酒后驾驶交通事故空间分析》.北京理工大学出版社,2022,114-115. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116862244A (en) | 2023-10-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110738135B (en) | Method and system for judging and guiding worker operation step standard visual recognition | |
| WO2019058379A1 (en) | Systems and methods for preventing work accidents | |
| CN110889339B (en) | Head and shoulder detection-based dangerous area grading early warning method and system | |
| Xu et al. | Dynamic safety prewarning mechanism of human–machine–environment using computer vision | |
| CN111428617A (en) | Video image-based distribution network violation maintenance behavior identification method and system | |
| CN114973140A (en) | Dangerous area personnel intrusion monitoring method and system based on machine vision | |
| CN110867046A (en) | Intelligent car washer video monitoring and early warning system based on cloud computing | |
| CN117035419B (en) | Intelligent management system and method for enterprise project implementation | |
| CN114782897A (en) | Dangerous behavior detection method and system based on machine vision and deep learning | |
| CN112163497B (en) | Construction site accident prediction method and device based on image recognition | |
| CN110023861A (en) | System and method for monitoring manufacturing works | |
| CN114862167A (en) | Production safety assessment management method and system for manufacturing production line | |
| CN113206978A (en) | Security intelligent monitoring early warning system and method for oil and gas pipeline station | |
| CN113506416A (en) | Engineering abnormity early warning method and system based on intelligent visual analysis | |
| CN117575334A (en) | Highway construction safety monitoring system | |
| CN115145788A (en) | Detection data generation method and device for intelligent operation and maintenance system | |
| CN116862244B (en) | Industrial field vision AI analysis and safety pre-warning system and method | |
| CN116503957B (en) | Gas household operation behavior identification method | |
| CN115346170B (en) | Intelligent monitoring method and device for gas facility area | |
| CN116909235A (en) | Intelligent factory monitoring method, intelligent factory monitoring system and electronic equipment | |
| CN103974028A (en) | Method for detecting fierce behavior of personnel | |
| Nain et al. | Authentication control system for the efficient detection of hard-hats using deep learning algorithms | |
| CN117172559B (en) | Risk identification early warning method, system and storage medium for Internet of things data | |
| CN117746338B (en) | Property park safety management method and system based on artificial intelligence | |
| CN116189101B (en) | Method and system for identifying, judging and guiding visual operation specification of security inspector |
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 | ||
| CB03 | Change of inventor or designer information |
Inventor after: Chen Dawei Inventor after: Zhong Hongliang Inventor after: Liang Shuting Inventor before: Zhong Hongliang Inventor before: Liang Shuting Inventor before: Hu Huilong |
|
| CB03 | Change of inventor or designer information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |