A smart factory safety supervision method under a multi-linkage safety management mechanism

Abstract

The application discloses a kind of intelligent factory safety supervision methods under multi-link safety control mechanism, it is related to intelligent factory safety supervision technical field, including the following steps, camera is arranged in the production field area of intelligent factory, and the monitoring video data of supervision area is collected by camera;Monitoring video data collected is segmented into multiple monitoring video segments based on equipment production running cycle time;The audio signal of each monitoring video segment after segmentation processing is collected, and the multiple monitoring video segments are analyzed based on audio signal Differentiation, obtain key monitoring video segment and feature time information;The selected key monitoring video segment is corrected based on feature time information and extracts key frame image;The application realizes to key feature monitoring video Frame extraction key frame image, significantly reduce the information redundancy of collection, avoid to appear multiple continuous frames repeatedly, enhance the content diversity of extracted key frame image, improve safety supervision accuracy and security.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent factory safety supervision, in particular to an intelligent factory safety supervision method under a multi-linkage safety management mechanism. BACKGROUND

[0002] An intelligent factory is based on intelligent technology, digital technology and information technology, and integrates personnel, machines, equipment and infrastructure in the factory through the fusion of physical infrastructure and information infrastructure to implement real-time management, coordination and control between multiple systems. On this basis, production is managed in a more refined and dynamic manner to achieve a "smart" state, thereby improving the management efficiency and production efficiency of the factory. Factory equipment safety supervision is a key link to ensure factory production safety and improve production efficiency. The main purpose of intelligent factory safety supervision is to monitor the running state of equipment in real time, timely discover and handle potential safety hazards, prevent production interruptions and safety accidents caused by equipment failure, and through monitoring, optimize equipment maintenance plans, extend equipment life, and reduce maintenance costs.

[0003] For example, the application publication no. CN116758484A, application publication date 2023.09.15, name "safety supervision system and method of chemical plant", extracts the production operation dynamic characteristics of the production operation monitoring video of the production workshop workers through a convolutional neural network model, and classifies the production operation dynamics to monitor and intelligently analyze the production operation of the workers. In this way, it can help the monitor to timely discover non-standard production operations, thereby reducing the occurrence of production safety accidents.

[0004] The existing technology including the above-mentioned application has the following deficiencies: the safety supervision of the intelligent factory in the prior art monitors the production operation of the factory through the equipment to collect many continuous frames of the production operation monitoring video, and when extracting images from the production operation monitoring video for analysis, the video is frame-extracted at a predetermined time interval to obtain a plurality of frame-extracted images of the video, and then feature recognition analysis is performed based on the collected frame-extracted images. This will cause information redundancy, significantly increase the amount of calculation for subsequent safety analysis and evaluation of the collected information, and affect the accuracy of intelligent factory safety supervision. SUMMARY

[0005] The purpose of the present application is to provide an intelligent factory safety supervision method under a multi-linkage safety management mechanism to solve the above-mentioned deficiencies in the prior art.

[0006] In order to achieve the above-mentioned purpose, the present application provides the following technical solution: an intelligent factory safety supervision method under a multi-linkage safety management mechanism, comprising the following steps:

[0007] Deploy cameras in the production area of ​​the smart factory to collect surveillance video data of the supervised area;

[0008] Segment the collected surveillance video data into multiple surveillance video segments based on the equipment production and operation cycle time;

[0009] Collect the audio signal of each surveillance video segment after segmentation, perform differential analysis on the intercepted multiple surveillance video segments based on the audio signal, and obtain key surveillance video segments and characteristic time information;

[0010] The selected key surveillance video segments are corrected based on the characteristic time information to extract key frame images, and the extracted key frame images are integrated to obtain a set of key surveillance video segment analysis images;

[0011] Analyze each image in the image set of key monitoring video segments to identify anomalies and determine whether there are images with abnormal movements. If there are abnormal movements, an alarm will be issued.

[0012] As a further description of the above technical solution: the equipment production operation cycle time is the time the equipment uses to produce and process a product. Each monitoring video segment includes a monitoring video of all operations of the relevant equipment in processing a product, ensuring the integrity of key information in the cropped and segmented monitoring video segments.

[0013] As a further description of the above technical solution: performing differential analysis on multiple intercepted surveillance video segments based on audio information to obtain key surveillance video segments and characteristic time information specifically includes the following steps:

[0014] Perform audio feature extraction on the audio signals collected from each surveillance video segment to obtain the noise time series diagram corresponding to each surveillance video segment;

[0015] The noise time series diagrams of each surveillance video segment are integrated into the same noise processing coordinate system, where the horizontal axis represents time and the vertical axis represents noise intensity;

[0016] The characteristic moment marking line perpendicular to the X-axis is drawn at the preset interval tj, and the intersection data information of the vertex and bottom points of the intersection of each characteristic moment marking line and the noise time series diagram of the noise processing coordinate system is collected.

[0017] The intersection data information includes:

[0018] T m (Z m1 ,Z m2 ) represents Z m1 Noise timing diagram and Z m2 Noise timing diagram and T m The intersection points of the characteristic moment marking lines are the apex and bottom points;

[0019] The key monitoring video segment and the feature time information are obtained based on intersection data information integration.

[0020] As a further description of the above technical solution: the key monitoring video segment and the feature time information are obtained based on intersection data information integration, specifically:

[0021] Based on the noise time sequence diagram included in the intersection data information, the monitoring video segment corresponding to the noise time sequence diagram is marked as a key monitoring video segment, and each feature time marker data corresponding to the noise time sequence diagram is marked as feature time information of the corresponding key monitoring video segment.

[0022] As a further description of the above technical solution: the selected key monitoring video segment is corrected based on the feature time information to extract key frame images, specifically including the following steps:

[0023] Collecting each feature time information of the key monitoring video segment, identifying and performing a correction process on the feature time information time sequence, and obtaining a feature time information set corresponding to the key monitoring video segment;

[0024] Performing a secondary correction process on the obtained feature time information set to obtain a frame extraction time information set corresponding to the key monitoring video segment;

[0025] Based on the frame extraction time information set corresponding to the key monitoring video segment, the key monitoring video segment is frame extracted to obtain a feature image set.

[0026] As a further description of the above technical solution: collecting each feature time information of the key monitoring video segment, identifying and performing a correction process on the feature time information time sequence, and obtaining a feature time information set corresponding to the key monitoring video segment specifically includes the following steps:

[0027] Traverse and analyze each feature time information in the image set, and when there are two adjacent feature time information in the same key monitoring video segment, calculate a corrected frame extraction time value, wherein the corrected frame extraction time value calculation method is: T i , T i+1 represents the two adjacent feature time information in the key monitoring video segment, T X represents T i , T i+1 the corrected frame extraction time information of the two adjacent feature time information;

[0028] Integrate to obtain a feature time information set corresponding to the key monitoring video segment.

[0029] As a further description of the above technical solution: the feature time information set is obtained by performing a secondary correction process on the obtained feature time information set, and a frame extraction time information set corresponding to the key monitoring video segment is obtained, specifically including the following steps:

[0030] retrieve the feature time information set obtained after the first correction processing;

[0031] mark a line on the noise time sequence graph corresponding to the key monitoring video segment based on each feature time information in the feature time information set at the corresponding marker feature moment;

[0032] extend F length distance to both sides with the feature moment marker line as the reference line to obtain a rectangular correction window, wherein F < tj, tj represents the preset interval time of the feature moment marker line;

[0033] When the feature moment marker line and the noise time sequence graph intersection are vertices, calculate the noise maximum value of the noise time sequence graph in the rectangular correction window, obtain the feature time information corresponding to the noise maximum value to replace the current feature time information and mark it as the frame extraction time information;

[0034] Process and obtain the frame extraction time information for each feature time information in the feature time information set in turn and integrate to obtain the frame extraction time information set corresponding to the key monitoring video segment.

[0035] As a further description of the above technical solution: based on the frame extraction time information set corresponding to the key monitoring video segment, the key monitoring video segment is processed to obtain a feature image set, specifically:

[0036] Retrieve each frame extraction time information in the frame extraction time information set, and extract the key frame image by extracting the key monitoring video segment according to the frame extraction time information;

[0037] Integrate the key frame images obtained by extracting the key monitoring video segment to obtain the feature image set of the key monitoring video segment.

[0038] As a further description of the above technical solution: the images in the key monitoring video segment analysis image set are subjected to abnormality recognition, and it is determined whether there is an abnormal motion image, and if there is an abnormal motion, an alarm is issued to remind Specifically:

[0039] Input each image in the key monitoring video segment analysis image set into the trained motion abnormality recognition model, and determine whether there is an abnormal motion image through the abnormality recognition model, and if there is an abnormal motion image, an alarm is issued to remind.

[0040] As a further description of the above technical solution: the training method of the abnormality recognition model is specifically:

[0041] Construct a motion abnormality recognition model based on a deep learning algorithm;

[0042] Collect a historical data set, the historical data set includes N training data sets, N is a positive integer, the training data set includes feature data and label data, the feature data is an analysis image set of a monitoring video segment, and the label data includes 1 and 0, wherein 1 represents an abnormal dynamic, and 0 represents no abnormal action;

[0043] Taking the feature data in each set of training data as the input of the action anomaly recognition model, taking the result of whether an abnormal action occurs predicted by the action anomaly recognition model for each set of feature data as the output, taking the label data corresponding to each set of feature data as the prediction target, and taking the sum of the minimum prediction accuracy as the training target, the action anomaly recognition model is trained until the sum of the prediction accuracy reaches convergence, and the training is stopped.

[0044] In the above technical solution, the present application provides a kind of intelligent factory safety supervision method under the multi-link security control mechanism, by extracting the audio signal of each monitoring video segment, obtaining noise time sequence diagram, and inserting feature time marker line based on preset interval time, collecting the intersection data information of the top point and the bottom point of the intersection point of feature time marker line and noise processing coordinate system noise time sequence diagram intersection point, and based on intersection data information, obtain key monitoring video segment and feature time information, to realize the frame extraction of key feature monitoring video, compared with the frame extraction image obtained by directly extracting frame according to preset interval time in prior art for analysis, can significantly reduce the information redundancy of collection, avoid the repetition of multiple continuous frames, enhance the content diversity of extracted key frame image, and then improve the accuracy and safety of intelligent factory safety supervision;

[0045] Secondly, by identifying the feature time information sequence, the adjacent two feature time information located in the same key monitoring video segment are integrated and extracted to obtain frame extraction time information, so that the same device abnormal feature corresponding key frame image can be reduced when collecting key frame image based on frame extraction time information subsequently, further reducing information redundancy, reducing analysis and calculation amount, and improving the accuracy of safety supervision. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0047] Figure 1 A flowchart of an intelligent factory safety supervision method under a multi-link security control mechanism is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0048] In order to make the technical personnel in the art better understand the technical solutions of the present application, the present application will be further described in detail below with reference to the drawings.

[0049] Please refer to Figure 1 The embodiment of the present application provides a technical scheme: a smart factory safety supervision method under a multi-link safety control mechanism, comprising the following steps:

[0050] Arranging a camera in the production site area of the smart factory, and collecting monitoring video data of the supervision area through the camera;

[0051] Segmenting the collected monitoring video data into multiple monitoring video segments based on the equipment production operation cycle time; the equipment production operation cycle time is the time used for the equipment to produce one product, so that each monitoring video segment obtained by segmenting the collected monitoring video data based on the equipment production operation cycle time includes monitoring video of all operations of the related equipment for processing one product, ensuring the integrity of the key information of the segmented monitoring video segment. At the same time, each monitoring video segment corresponds consistently in time sequence, thereby facilitating subsequent analysis and processing of each collected monitoring video segment.

[0052] Collecting the audio signal of each monitoring video segment after segmentation and processing, and performing differential analysis on the intercepted multiple monitoring video segments based on the audio signal to obtain key monitoring video segments and feature time information;

[0053] Correcting the selected key monitoring video segments based on the feature time information to extract key frame images, and integrating and processing the extracted key frame images to obtain a key monitoring video segment analysis image set;

[0054] Abnormality recognition is performed on each image in the key monitoring video segment analysis image set to determine whether there is an image with abnormal motion, and an alarm is issued if there is an abnormal motion. Each image in the key monitoring video segment analysis image set is input into the trained motion abnormality recognition model, and the abnormality recognition model is used to determine whether there is an image with abnormal motion, and an alarm is issued if there is an image with abnormal motion.

[0055] In another embodiment of the present application, differential analysis is performed on the intercepted multiple monitoring video segments based on audio information to obtain key monitoring video segments and feature time information, which specifically comprises the following steps:

[0056] Audio feature extraction is performed on the audio signal collected by each monitoring video segment to obtain a noise time sequence diagram corresponding to each monitoring video segment; the noise time sequence diagram is a line graph representing the change of noise intensity with time as the sequence;

[0057] The noise time sequence diagrams of the monitoring video segments are integrated in the same noise processing coordinate system, wherein the horizontal axis represents time and the vertical axis represents noise intensity.

[0058] The preset interval time tj is used as a characteristic time mark line perpendicular to the X axis, intersection data information of the top and bottom points of the intersection of each characteristic time mark line and the noise time sequence diagram of the noise processing coordinate system is collected,

[0059] The intersection data information includes:

[0060] T m (Z m1 ,Z m2 ) represents Z m1 The noise time sequence diagram and Z m2 The noise time sequence diagram and Z m The intersection points of the characteristic time mark line are the top and bottom points; wherein the top and bottom points of the intersection points of the characteristic time mark lines at different times correspond to the same noise time sequence diagram, and the top point of the intersection point of the characteristic time mark line and the noise time sequence diagram of the noise processing coordinate system represents that the noise intensity of the monitoring video segment corresponding to the top point of the time node of each monitoring video segment is the maximum, and the bottom point of the intersection point of the characteristic time mark line and the noise time sequence diagram of the noise processing coordinate system represents that the noise intensity of the monitoring video segment corresponding to the top point of the time node of each monitoring video segment is the minimum.

[0061] Based on the intersection data information, the key monitoring video segment and the characteristic time information are obtained. That is, the audio signal of each monitoring video segment is extracted, the noise time sequence diagram is obtained, the characteristic time mark line is inserted based on the preset interval time, the intersection data information of the top and bottom points of the intersection of the characteristic time mark line and the noise time sequence diagram of the noise processing coordinate system is collected, and the key monitoring video segment and the characteristic time information are obtained based on the intersection data information, so that the key frame image is extracted by frame extraction of the key characteristic monitoring video. Compared with the prior art of directly extracting the frame image by frame extraction of the video according to the preset interval time, the information redundancy of the collected information can be significantly reduced, the repetition of multiple continuous frames can be avoided, the content diversity of the extracted key frame image can be enhanced, and the accuracy and safety of the intelligent factory safety supervision can be improved.

[0062] In still another embodiment of the application, based on the intersection data information, the key monitoring video segment and the characteristic time information are obtained, specifically:

[0063] Based on the intersection data information, the noise time sequence diagram is marked as the key monitoring video segment, and the characteristic time mark line data corresponding to the noise time sequence diagram is marked as the characteristic time information of the corresponding key monitoring video segment.

[0064] In still another embodiment of the application, the selected key monitoring video segment is corrected based on the characteristic time information to extract the key frame image, specifically including the following steps:

[0065] collecting each feature time information of the key monitoring video segment, identifying the time sequence of the feature time information and performing a first correction processing to obtain a feature time information set corresponding to the key monitoring video segment;

[0066] performing a second correction processing on the obtained feature time information set to obtain a frame extraction time information set corresponding to the key monitoring video segment;

[0067] performing frame extraction processing on the key monitoring video segment based on the frame extraction time information set corresponding to the key monitoring video segment to obtain a feature image set.

[0068] In still another embodiment of the application, collecting each feature time information of the key monitoring video segment, identifying the time sequence of the feature time information and performing a first correction processing to obtain a feature time information set corresponding to the key monitoring video segment specifically includes the following steps:

[0069] traversing and analyzing each feature time information in the image set, and calculating a corrected frame extraction time value when there are two adjacent feature time information in the same key monitoring video segment, wherein the method for calculating the corrected frame extraction time value is: T i , T i+1 denote the two adjacent feature time information in the key monitoring video segment, T X denotes T i , T i+1 the corrected frame extraction time information of the two adjacent feature time information;

[0070] integrating the obtained feature time information set corresponding to the key monitoring video segment. It should be noted that when the two adjacent frames are in the same key monitoring video segment, the device abnormal features corresponding to the two adjacent frames are obviously consistent. The application identifies the time sequence of the feature time information, integrates and extracts the frame extraction time information of the two adjacent feature time information in the same key monitoring video segment, and reduces the key frame images corresponding to the same device abnormal features when collecting the key frame images based on the frame extraction time information subsequently, further reduces information redundancy, reduces analysis and calculation amount, and improves the accuracy of safety supervision.

[0071] In still another embodiment of the application, the second correction processing on the obtained feature time information set to obtain a frame extraction time information set corresponding to the key monitoring video segment specifically includes the following steps:

[0072] calling the feature time information set obtained after the first correction processing;

[0073] marking a feature time marker line on a noise time sequence graph of the corresponding key monitoring video segment based on each feature time information in the feature time information set;

[0074] Taking the feature time marker line as a reference line, a rectangular correction window is obtained by extending to both sides by a distance F, wherein F < tj, and tj represents a preset interval time of the feature time marker line;

[0075] When the intersection point of the feature time marker line and the noise time sequence diagram is a vertex, the maximum noise value of the noise time sequence diagram in the rectangular correction window is calculated, the feature time information corresponding to the maximum noise value is obtained, the current feature time information is replaced, and the frame extraction time information is marked; it should be noted that when the intersection point of the feature time marker line and the noise time sequence diagram is a bottom point, the minimum noise value of the noise time sequence diagram in the rectangular correction window is calculated, the feature time information corresponding to the minimum noise value is obtained, the current feature time information is replaced, and the frame extraction time information is marked;

[0076] The feature time information in the feature time information set is processed and the frame extraction time information is obtained, and the frame extraction time information set corresponding to the key monitoring video segment is obtained.

[0077] By inserting the feature time marker line, the intersection data information of the top and bottom points of the intersection point of the feature time marker line and the noise processing coordinate system noise time sequence diagram is collected, and the key monitoring video segment and the feature time information are obtained based on the intersection data information, at this time, when the frame extraction is collected by directly obtaining the feature time information, part of the important feature information will be missed due to the preset interval time, and the application modifies the feature time information determined by each feature time marker line to obtain the frame extraction time information, so that the extracted key frame image feature is more prominent, and the accuracy of subsequent judgment of whether there is motion abnormality is further improved.

[0078] In still another embodiment of the application, the key monitoring video segment is processed by frame extraction based on the frame extraction time information set corresponding to the key monitoring video segment to obtain a feature image set, specifically:

[0079] Each frame extraction time information in the frame extraction time information set is called, and the key monitoring video segment is frame extracted according to the frame extraction time information, and the obtained frame extraction image is marked as a key frame image.

[0080] The key frame images obtained by frame extraction of the key monitoring video segment are integrated to obtain a feature image set of the key monitoring video segment.

[0081] The training method of the abnormality recognition model is specifically:

[0082] An action abnormality recognition model is constructed based on a deep learning algorithm;

[0083] Collect a historical data set, the historical data set includes N training data sets, N is a positive integer, the training data set includes feature data and label data, the feature data is an image set analyzed by a monitoring video segment, and the label data includes 1 and 0, wherein 1 represents an abnormal dynamic, and 0 represents no abnormal action; the feature data in each set of training data is used as input of an action anomaly recognition model, whether an abnormal action occurs is predicted by the action anomaly recognition model as output of each set of feature data, the label data corresponding to each set of feature data is used as a prediction target, and the action anomaly recognition model is trained with the sum of prediction accuracies as a training target until the sum of prediction accuracies reaches convergence, and the training is stopped. Wherein the training of the action anomaly recognition model is prior art.

[0084] In the description of the present specification, the description of the terms "one embodiment", "example", "specific example" and the like means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0085] The preferred embodiments of the application disclosed above are only used to help explain the application. The preferred embodiments do not describe all the details and limit the application to the specific embodiments described. Obviously, many modifications and changes can be made according to the content of the present specification. The present specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the application, so that those skilled in the art can well understand and utilize the application. The application is limited only by the claims and their full scope and equivalents.

Claims

1. A smart factory safety supervision method under a multi-link safety management mechanism, characterized in that, The method comprises the following steps: arranging a camera in a smart factory production site area, and collecting monitoring video data of the supervision area through the camera; segmenting the collected monitoring video data into multiple monitoring video segments based on the equipment production operation cycle time; collecting audio signals of each monitoring video segment after segmentation processing, and performing differential analysis on the intercepted multiple monitoring video segments based on the audio signals to obtain key monitoring video segments and feature time information; Specifically, the method comprises the following steps: extracting audio features from the audio signals collected for each monitoring video segment to obtain a noise timing diagram corresponding to each monitoring video segment; integrate the noise timing diagrams of each monitoring video segment in the same noise processing coordinate system, where the horizontal axis represents time and the vertical axis represents noise intensity; preset interval time tj as a feature time marker line perpendicular to the X-axis, collect the intersection data information of the top and bottom points of each feature time marker line and the intersection point of the noise processing coordinate system noise timing diagram, The intersection data information includes: ; representing noise timing diagram and noise timing diagram and the intersection of the feature time marker line is the vertex and the bottom point; integrate the intersection data information to obtain key monitoring video segments and feature time information; correct the selected key monitoring video segments based on the feature time information to extract key frame images, and integrate and process the extracted key frame images to obtain a key monitoring video segment analysis image set; perform abnormality recognition on each image in the key monitoring video segment analysis image set to determine whether there is an image with abnormal motion, and issue an alarm if there is an abnormal motion. 2.The smart factory safety supervision method under a multi-action safety control mechanism according to claim 1, wherein, The equipment production operation cycle time is the time used for the production and processing of one product by the equipment, and each monitoring video segment includes the monitoring video of all operations of the related equipment for processing one product, ensuring the integrity of the key information of the cut and segmented monitoring video segments. 3.The smart factory safety supervision method under a multi-action safety control mechanism according to claim 1, wherein, Integrating the intersection data information to obtain key monitoring video segments and feature time information specifically comprises: Based on the intersection data information including each noise timing diagram, the monitoring video segment corresponding to the noise timing diagram is marked as a key monitoring video segment, and the feature time marker line data corresponding to the noise timing diagram is marked as the feature time information of the corresponding key monitoring video segment.

4. The intelligent factory safety monitoring method under the multi- linkage safety management mechanism according to claim 3, characterized in that, Correcting the selected key monitoring video segments based on the feature time information to extract key frame images specifically comprises the following steps: Collecting each feature time information of the key monitoring video segment, performing a one-time correction processing on the feature time information sequence, and obtaining a feature time information set corresponding to the key monitoring video segment; performing a second correction processing on the obtained feature time information set to obtain an extraction time information set corresponding to the key monitoring video segment; performing extraction processing on the key monitoring video segment based on the extraction time information set corresponding to the key monitoring video segment to obtain a feature image set.

5. The intelligent factory safety monitoring method under the multi- linkage safety management mechanism according to claim 4, characterized in that, Collecting each feature time information of the key monitoring video segment, performing a one-time correction processing on the feature time information sequence, and obtaining a feature time information set corresponding to the key monitoring video segment specifically comprises the following steps: The feature time information in the image set is traversed, and when there are two adjacent feature time information in the same key monitoring video segment, a modified frame extraction time value is calculated, wherein the calculation method of the modified frame extraction time value is: , 、 , , 、 the modified frame extraction time information of the two adjacent feature time information. Integrate the feature time information set corresponding to the key monitoring video segment.

6. The intelligent factory safety monitoring method under the multi- linkage safety management mechanism according to claim 5, characterized in that, The second correction processing on the obtained feature time information set to obtain an extraction time information set corresponding to the key monitoring video segment specifically comprises the following steps: retrieve the feature time information set obtained after the one-time correction processing; The feature time information in the feature time information set is marked on the noise time sequence diagram of the corresponding key monitoring video segment. A rectangular correction window is obtained by extending the feature time marker line by a distance F on both sides, where F < tj, and tj represents the preset interval time of the feature time marker line. When the intersection of the feature time marker line and the noise time sequence diagram is the vertex, the maximum noise value in the rectangular correction window is calculated, and the feature time information corresponding to the maximum noise value is obtained to replace the current feature time information and is marked as the frame extraction time information. The feature time information set corresponding to the key monitoring video segment is obtained by processing and integrating the frame extraction time information obtained from each feature time information in the feature time information set.

7. The intelligent factory safety monitoring method under the multi- linkage safety management mechanism according to claim 6, characterized in that, The feature image set is obtained by extracting the frame extraction time information set and performing frame extraction on the key monitoring video segment according to the frame extraction time information. The feature image set of the key monitoring video segment is obtained by integrating the key frame images obtained by frame extraction of the key monitoring video segment. The images in the key monitoring video segment analysis image set are identified for abnormality to determine whether there are images with abnormal motion, and if there are abnormal motions, an alarm is issued to remind the user.

8. The intelligent factory safety monitoring method under the multi- linkage safety management mechanism according to claim 1, characterized in that, The images in the key monitoring video segment analysis image set are identified for abnormality to determine whether there are images with abnormal motion, and if there are abnormal motions, an alarm is issued to remind the user. The training method of the abnormality recognition model is as follows: 9.The smart factory safety supervision method under a multi-action safety control mechanism according to claim 8, wherein, An action abnormality recognition model is constructed based on a deep learning algorithm. A historical data set is collected, which includes N training data sets, where N is a positive integer, and the training data set includes feature data and label data, the feature data is a monitoring video segment analysis image set, and the label data includes 1 and 0, where 1 represents an abnormal motion and 0 represents no abnormal motion. The feature data in each training data set is used as the input of the action abnormality recognition model, the result of the action abnormality recognition model predicting whether there is an abnormal motion for each feature data is used as the output, the label data corresponding to each feature data is used as the prediction target, and the action abnormality recognition model is trained to minimize the sum of prediction accuracy until the sum of prediction accuracy converges. ​

Images