A Workplace Safety Analysis Method Based on Personnel Behavior Analysis

Abstract

This invention relates to the field of work safety management technology, and provides a work site safety analysis method based on personnel behavior analysis. The method includes receiving behavior monitoring requests and parsing worker information, acquiring behavior data and generating behavior feature information, displaying the feature information and risk assessment controls through an analysis pop-up window, and performing risk assessment based on configured risk levels. When an abnormal behavior request is detected, verification rules are generated based on a preset behavior library and sent to the safety monitoring device. The rules are displayed on an external monitor, and the device receives the behavior data to be verified. Noise filtering and feature extraction are performed on the data to be verified to generate behavior feature vectors. Similarity is calculated with historical feature vectors, and target similarities that meet a threshold are selected to determine matching feature vectors. If a match is successful, a safety warning is triggered. This method covers the entire process of data acquisition, feature processing, risk matching, and warning.

Description

Technical Field

[0001] This invention relates to the field of work safety management technology, and more specifically, to a work site safety analysis method based on personnel behavior analysis. Background Technology

[0002] In modern industrial production, on-site safety management has always been a crucial and highly regarded aspect. With the continuous expansion of production scale and the increasing complexity of working environments, the impact of worker behavior on on-site safety is becoming increasingly significant. Traditional on-site safety management primarily relies on manual inspections and simple monitoring equipment, its technical principle mainly involving manual observation or video playback to identify potential safety hazards. However, this approach is not only inefficient but also susceptible to human factors, making it difficult to identify abnormal worker behavior in real time and accurately. In recent years, although some sensor-based or video analytics technologies have been introduced into safety management, these technologies mostly only enable single-type behavior monitoring and have shortcomings in behavioral feature extraction and risk assessment, failing to meet the high-precision and real-time requirements of safety analysis in complex work environments.

[0003] In implementing the embodiments of the present invention, the inventors discovered at least the following problems or defects in the prior art: First, traditional safety management methods cannot achieve real-time monitoring and dynamic analysis of worker behavior, making it difficult to detect potential safety hazards in a timely manner; second, when extracting behavioral features, the prior art often ignores the completeness and accuracy of behavioral data, resulting in unreliable analysis results; third, for different types of behavioral monitoring requests, the prior art lacks a flexible processing mechanism and cannot meet diverse behavioral analysis needs; finally, the prior art mostly adopts fixed rules in risk assessment, lacking the ability to deeply mine and dynamically adjust behavioral feature information, making it difficult to adapt to complex and ever-changing work site environments. Summary of the Invention

[0004] This invention provides a method, equipment, and medium for workplace safety analysis based on personnel behavior analysis.

[0005] In a first aspect of the present invention, a workplace safety analysis method based on personnel behavior analysis is provided, comprising:

[0006] In response to receiving a behavior monitoring request through a target application associated with a work site safety monitoring device, the behavior monitoring request is parsed to obtain the worker information corresponding to the behavior monitoring request, wherein the behavior monitoring request is a request to perform behavior monitoring on the current work site;

[0007] Obtain the behavioral data of the workers;

[0008] Based on the behavioral data, behavioral feature information is generated;

[0009] An analysis pop-up window corresponding to the behavioral characteristic information is displayed, wherein the behavioral characteristic information and risk assessment configuration controls are displayed in the analysis pop-up window;

[0010] Based on the risk level information corresponding to the configured risk assessment control, the behavioral characteristic information is assessed for risk.

[0011] In response to an abnormal behavior request detected by the on-site safety monitoring device, a behavior verification rule is generated based on a preset behavior library;

[0012] The behavior verification rules are sent to the work site safety monitoring device, so that the work site safety monitoring device displays the behavior verification rules on an external display.

[0013] The system receives the behavior data to be verified, which corresponds to the behavior verification rules, sent by the on-site safety monitoring device as the information of the person to be identified. The behavior data to be verified is the behavior record of the worker during the monitoring period.

[0014] In response to obtaining the information of the person to be identified through the work site safety monitoring device, the information of the person to be identified is matched according to the stored behavioral feature information to obtain a matching result. The matching process of the information of the person to be identified according to the stored behavioral feature information to obtain a matching result includes: performing noise filtering and invalid data removal on the behavioral data to be verified to obtain processed behavioral data.

[0015] The processed behavioral data is subjected to feature extraction to obtain a behavioral feature vector;

[0016] Generate the behavioral similarity between the behavioral feature vector and the historical feature vectors included in each behavioral feature information, and obtain a similarity set;

[0017] The similarity scores in the similarity set that satisfy a first preset similarity condition are determined as the target similarity scores, wherein the first preset similarity condition is that the similarity score is the maximum value in the similarity set;

[0018] In response to determining that the target similarity satisfies a second preset similarity condition, the feature vector corresponding to the target similarity is determined as a matching feature vector, wherein the second preset similarity condition is that the target similarity is greater than a preset similarity threshold;

[0019] Based on the personnel information corresponding to the matching feature vector, a matching result is generated;

[0020] In response to determining that the matching result indicates a successful match, the early warning component of the on-site safety monitoring device is controlled to perform a safety early warning operation.

[0021] Furthermore, the worker information includes worker identification and location information; and before acquiring the worker's behavioral data, the method further includes:

[0022] In response to detecting a selection operation of the launch control on the behavior monitoring page, the location information of the operator corresponding to the behavior monitoring request is obtained according to the operator identifier, wherein the behavior monitoring page corresponds to the behavior monitoring request, and the operator information, including the operator's name, is displayed on the behavior monitoring page;

[0023] Determine whether the location information meets the preset range conditions corresponding to the target work area;

[0024] In response to determining that the location information meets the preset range condition, a behavior data collection instruction corresponding to the behavior monitoring request is sent to the work site safety monitoring device; and the acquisition of the behavior data of the worker includes: in response to receiving feedback information indicating consent to collection corresponding to the behavior data collection instruction, acquiring the behavior data of the worker.

[0025] Further, generating behavioral feature information based on the behavioral data includes:

[0026] The behavioral data is subjected to integrity detection processing to obtain integrity detection results;

[0027] In response to determining that the integrity detection result indicates the existence of a valid behavior record, behavior feature information is generated based on the behavior data.

[0028] Furthermore, the behavior monitoring request is a video behavior monitoring request, and the behavior data is behavior video data; and the process of performing integrity detection processing on the behavior data to obtain an integrity detection result includes: performing action recognition processing on the behavior data to obtain an action recognition result as the integrity detection result.

[0029] Further, generating behavioral feature information based on the behavioral data includes:

[0030] The behavioral data is segmented into frames to obtain a behavioral data sequence;

[0031] For each behavioral data unit in the behavioral data sequence, perform the following steps: determine whether a complete human outline region exists in the behavioral data unit;

[0032] In response to determining that a complete human contour region exists in the behavior data unit, the human contour region in the behavior data unit is cropped to obtain contour data.

[0033] The contour data is color simplified to obtain simplified contour data;

[0034] Brightness detection is performed on the simplified contour data to obtain contour data intensity information;

[0035] Contrast detection is performed on the simplified contour data to obtain contour data comparison information;

[0036] Based on the simplified contour data quality information, the contour data intensity information, and the contour data comparison information, contour data parameters are generated.

[0037] Based on the obtained contour data parameters, select the contour data that meets the preset data parameter conditions from the obtained contour data as the target contour data;

[0038] Based on the simplified contour data quality information, contour data intensity information, and contour data comparison information corresponding to the target contour data, the target contour data is optimized to obtain optimized contour data.

[0039] The optimized contour data is subjected to feature extraction processing to obtain contour data feature information;

[0040] Behavioral feature information is generated based on the contour data feature information and the personnel information.

[0041] Further, the step of performing color simplification processing on the contour data to obtain simplified contour data includes:

[0042] The contour data is color simplified to obtain simplified contour data;

[0043] Remove each data point that meets the preset edge conditions from the simplified contour data to obtain the removed contour data;

[0044] For each data point in the simplified contour data, perform the following steps: determine the data point that is right-adjacent to the data point in the simplified contour data as the right-adjacent data point;

[0045] In the simplified contour data, the data points that are adjacent to the data points are defined as the upper adjacent data points;

[0046] Generate a first value and a second value based on the value of the right adjacent data point and the value of the data point;

[0047] Based on the values ​​of the adjacent data points and the values ​​of the data points, a third value and a fourth value are generated;

[0048] The sum of the first, second, third, and fourth values ​​is determined as the fifth value.

[0049] The number of each data point included in the simplified contour data is determined as the total number of data points;

[0050] The ratio of the fifth value to the total number of data points is determined as the contour data quality information corresponding to the simplified contour data.

[0051] Further, the step of performing data optimization processing on the target contour data based on the simplified contour data quality information, contour data intensity information, and contour data comparison information corresponding to the target contour data to obtain optimized contour data includes:

[0052] In response to determining that the quality information of the simplified contour data corresponding to the target contour data meets the preset quality optimization conditions, the target contour data is subjected to quality enhancement processing to update the target contour data;

[0053] In response to determining that the contour data intensity information corresponding to the target contour data meets the preset intensity optimization conditions, the target contour data is subjected to intensity adjustment processing to update the target contour data;

[0054] In response to determining that the contour data comparison information corresponding to the target contour data meets the preset comparison optimization conditions, the target contour data is subjected to comparison enhancement processing to update the target contour data;

[0055] The updated target contour data is then normalized in size to update the target contour data.

[0056] The updated target contour data is used as the optimized contour data.

[0057] Further, the behavior monitoring request is an action behavior monitoring request, and the behavior data is motion sensor data; and the process of performing integrity detection processing on the behavior data to obtain an integrity detection result includes: performing action feature extraction processing on the behavior data to obtain action feature results as integrity detection results; and the process of generating behavior feature information based on the behavior data includes:

[0058] The behavioral data is subjected to noise filtering and invalid data removal to obtain processed data;

[0059] The processed data is subjected to feature extraction to obtain an action feature vector;

[0060] For each feature dimension in the action feature vector, perform the following steps: determine the feature value corresponding to the feature dimension;

[0061] Determine the feature weights corresponding to the feature dimensions;

[0062] A comprehensive feature score is generated based on the feature values ​​and feature weights;

[0063] Select the comprehensive feature score that meets the preset scoring conditions from the various generated comprehensive feature scores as the target comprehensive feature score;

[0064] The target comprehensive feature score and the personnel information are determined as behavioral feature information.

[0065] In a second aspect of the invention, an electronic device is provided, comprising: at least one processor, a memory, and an input / output unit; wherein the memory is used to store a computer program, and the processor is used to invoke the computer program stored in the memory to perform the method described in any one of the first aspects.

[0066] In a third aspect of the invention, a computer-readable storage medium is provided, comprising instructions that, when executed on a computer, cause the computer to perform the method described in any one of the first aspects.

[0067] The embodiments of the present invention have at least the following beneficial effects: This work site safety analysis method based on personnel behavior analysis can effectively improve the safety of the work site. By real-time monitoring and analysis of the behavior data of workers, combined with the generation of behavioral feature information and the flexible application of risk assessment configuration controls, abnormal behavior can be detected in a timely manner and accurate risk assessment can be performed. When abnormal behavior is detected, behavior verification rules are generated based on a preset behavior library, and the behavior data of workers is matched and processed, thereby achieving accurate identification and early warning, effectively preventing potential safety accidents, and ensuring the personal safety of workers and the normal operation of the work site. In addition, this method can also flexibly process behavior data according to different types of behavior monitoring requests, such as video behavior monitoring requests and motion behavior monitoring requests, further improving the accuracy and reliability of behavior analysis and providing strong support for work site safety management.

[0068] Furthermore, this method can optimize the efficiency of safety management at work sites. Before acquiring worker behavior data, verifying whether the workers' location information meets preset range conditions ensures the targeted and effective collection of behavior data, avoids interference from irrelevant data, and improves data processing efficiency. When generating behavioral feature information, the method performs integrity checks, noise filtering, invalid data removal, and feature extraction on the behavior data, ensuring the quality and accuracy of the behavioral feature information and providing a reliable basis for subsequent risk assessment and behavior matching. Simultaneously, this method further enhances the usability of behavioral feature information through data optimization processes such as quality enhancement, intensity adjustment, and contrast enhancement, making the entire safety analysis process more efficient and accurate. This helps reduce the labor and time costs of work site safety management and improves the level of intelligent safety management. Attached Figure Description

[0069] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of the invention are illustrated in the drawings by way of example and not limitation, wherein:

[0070] Figure 1 This is a flowchart illustrating a work site safety analysis method based on personnel behavior analysis, provided as an embodiment of the present invention. Detailed Implementation

[0071] The principles and spirit of the invention will now be described with reference to several exemplary embodiments. It should be understood that these embodiments are provided merely to enable those skilled in the art to better understand and implement the invention, and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided to make the invention more thorough and complete, and to fully convey the scope of the invention to those skilled in the art.

[0072] Those skilled in the art will understand that embodiments of the present invention can be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0073] It should be noted that the number of any elements in the accompanying drawings is for illustrative purposes only and not as a limitation, and any naming is for distinction only and has no limiting meaning.

[0074] The following is for reference. Figure 1 , Figure 1 This is a flowchart illustrating a workplace safety analysis method based on personnel behavior analysis, provided as an embodiment of the present invention. Figure 1As shown, a workplace safety analysis method based on personnel behavior analysis includes:

[0075] S1 responds to receiving a behavior monitoring request through a target application associated with the work site safety monitoring device, parses the behavior monitoring request, and obtains the worker information corresponding to the behavior monitoring request, wherein the behavior monitoring request is a request to perform behavior monitoring on the current work site;

[0076] S2 acquires the behavioral data of the operators;

[0077] S3 generates behavioral feature information based on the behavioral data;

[0078] S4 displays an analysis pop-up window corresponding to the behavioral characteristic information, wherein the analysis pop-up window displays the behavioral characteristic information and risk assessment configuration controls;

[0079] S5 performs a risk assessment on the behavioral characteristic information based on the risk level information corresponding to the configured risk assessment configuration control;

[0080] S6 responds to the abnormal behavior request detected by the work site safety monitoring device by generating behavior verification rules according to the preset behavior library;

[0081] S7 sends the behavior verification rules to the work site safety monitoring device, so that the work site safety monitoring device displays the behavior verification rules through an external display;

[0082] S8 receives the behavior data to be verified corresponding to the behavior verification rule sent by the work site safety monitoring device as the information of the person to be identified, wherein the behavior data to be verified is the behavior record of the worker during the monitoring period;

[0083] S9 responds to obtaining the information of the person to be identified through the work site safety monitoring device, and performs matching processing on the information of the person to be identified according to the stored behavioral feature information to obtain a matching result. The matching processing on the information of the person to be identified according to the stored behavioral feature information to obtain a matching result includes: performing noise filtering and invalid data removal processing on the behavioral data to be verified to obtain processed behavioral data.

[0084] S10 performs feature extraction processing on the processed behavior data to obtain a behavior feature vector;

[0085] S11 generates the behavioral similarity between the behavioral feature vector and the historical feature vectors included in each behavioral feature information, and obtains a similarity set;

[0086] S12 determines the similarity in the similarity set that satisfies the first preset similarity condition as the target similarity, wherein the first preset similarity condition is that the similarity is the maximum value in the similarity set;

[0087] S13 In response to determining that the target similarity satisfies the second preset similarity condition, the feature vector corresponding to the target similarity is determined as the matching feature vector, wherein the second preset similarity condition is that the target similarity is greater than a preset similarity threshold;

[0088] S14 generates a matching result based on the personnel information corresponding to the matching feature vector;

[0089] S15 responds to the determination that the matching result indicates a successful match by controlling the early warning component of the on-site safety monitoring device to perform a safety early warning operation.

[0090] It should be noted that this invention proposes a workplace safety analysis method based on personnel behavior analysis. Here, a behavior monitoring request refers to a request initiated by a target application associated with a workplace safety monitoring device to request behavior monitoring of the current workplace. This request includes personnel information, such as personnel identification and location, used to determine the objects and scope to be monitored. Behavioral data refers to records of personnel behavior at the workplace, which may include video data, motion sensor data, etc., and this data forms the basis for subsequent analysis. By parsing the behavior monitoring request to obtain personnel information and further acquiring behavioral data, a basis can be provided for subsequent generation of behavioral characteristic information and risk assessment. The core of this method lies in the timely detection and early warning of potential safety hazards through real-time monitoring and analysis of personnel behavior.

[0091] Specifically, the behavior monitoring request mentioned in this invention is a key concept. It is initiated by the safety monitoring system at the work site to start the behavior monitoring process for specific workers. Worker information includes the worker's identification and location information. The identification can be an employee ID, name, or other unique identifier used to distinguish different workers; the location information can be the worker's specific coordinates at the work site, obtained through positioning technology, used to determine whether the worker is within the area to be monitored. Behavioral data refers to various records of worker behavior during the work process, such as behavioral video data collected by video surveillance equipment or motion sensor data collected by motion sensors. This data undergoes a series of processing steps after collection, such as integrity checks and noise filtering, to ensure data quality and usability. Furthermore, the risk assessment configuration control is a user interface component used to set and adjust risk assessment parameters and rules, such as risk level classification criteria. Through these configurations, the system can flexibly conduct risk assessments according to different work site requirements and safety standards.

[0092] Preferably, the processing of behavioral data can be further refined. Taking behavioral video data as an example, the video data is first processed by frame segmentation, that is, the video is decomposed into a series of frame images to form a sequence of behavioral data. For each frame image, the system detects whether a complete personnel outline region exists. If it exists, the region is cropped to obtain outline data. Next, the outline data undergoes color simplification processing, for example, simplifying complex color information into grayscale images or other limited color ranges to reduce data volume and highlight key features. At the same time, brightness and contrast detection are performed on the simplified outline data to obtain outline data intensity information and contrast information, respectively. This information is used to generate outline data parameters, and then select outline data that meets the preset data parameter conditions as target outline data. For the target outline data, data optimization processing is also performed, such as quality enhancement, intensity adjustment, and contrast enhancement, to further improve the quality and usability of the data. Finally, feature extraction processing is used to obtain outline data feature information, and combined with personnel information to generate behavioral feature information. This behavioral feature information will serve as an important basis for subsequent risk assessment, thereby achieving accurate analysis and risk warning of worker behavior.

[0093] In some embodiments, the worker information includes worker identification and location information; and before acquiring the worker's behavior data, the method further includes:

[0094] In response to detecting a selection operation of the launch control on the behavior monitoring page, the location information of the operator corresponding to the behavior monitoring request is obtained according to the operator identifier, wherein the behavior monitoring page corresponds to the behavior monitoring request, and the operator information, including the operator's name, is displayed on the behavior monitoring page;

[0095] Determine whether the location information meets the preset range conditions corresponding to the target work area;

[0096] In response to determining that the location information meets the preset range condition, a behavior data collection instruction corresponding to the behavior monitoring request is sent to the work site safety monitoring device; and the acquisition of the behavior data of the worker includes: in response to receiving feedback information indicating consent to collection corresponding to the behavior data collection instruction, acquiring the behavior data of the worker.

[0097] It should be noted that before acquiring the worker's behavioral data, this invention detects the selection operation of the start control on the behavior monitoring page, obtains the worker's location information based on their identifier, and determines whether this location information meets the preset range conditions of the target work area. This process ensures the relevance and effectiveness of behavioral data collection, avoiding unnecessary data collection from workers outside the target area. Only when the worker's location meets the preset range conditions will a behavioral data collection command be sent to the on-site safety monitoring device, thus initiating the behavioral data collection process. This method improves data collection efficiency while reducing interference from invalid data, providing more accurate basic data for subsequent behavioral analysis.

[0098] Specifically, the behavior monitoring page is the user interface associated with the behavior monitoring request, used to display worker information, such as worker name. The launch control is an operation button on this page; when a user (such as a safety manager or system operator) clicks this control, the system triggers the subsequent worker information acquisition process. The worker identifier is an identifier used to uniquely identify the worker; it can be an employee ID, national ID number, or other unique code. Location information refers to the worker's specific location coordinates at the work site, obtained through positioning technologies (such as GPS, Wi-Fi positioning, or RFID). Preset range conditions refer to the boundary conditions of the specific area in the work site where behavior monitoring needs to be performed, such as the coordinate range of a specific work area. When the worker's location information falls within this range, the system considers the worker to be within the target work area, thus triggering the sending of a behavior data collection command. This command is a signal generated by the system to notify the work site safety monitoring device to begin collecting behavior data. Feedback information is the signal returned by the monitoring device after receiving the collection command, indicating that it is ready and agrees to begin collecting behavior data.

[0099] Preferably, the acquisition of location information and the setting of preset range conditions can be further refined. For example, location information can be acquired in real time through positioning tags (such as RFID tags or smart safety helmets) installed on workers, which periodically send their location coordinates to the system. The preset range conditions can be a polygonal area enclosed by multiple coordinate points, or a circular area with a certain radius centered on a fixed point. The system will determine whether the worker's position meets the conditions based on these coordinate ranges. When sending behavioral data collection instructions, parameters such as instruction priority and timeout can be set to ensure that the instructions are received and responded to by the monitoring device in a timely and effective manner. For example, if the monitoring device does not return feedback information within a specified time, the system can automatically resend the instruction or trigger an alarm to ensure the smooth progress of behavioral data collection.

[0100] In some embodiments, generating behavioral feature information based on the behavioral data includes:

[0101] The behavioral data is subjected to integrity detection processing to obtain integrity detection results;

[0102] In response to determining that the integrity detection result indicates the existence of a valid behavior record, behavior feature information is generated based on the behavior data.

[0103] It should be noted that before generating behavioral feature information, this invention performs an integrity check on the behavioral data to ensure the validity and usability of the collected data. The integrity check result determines whether to continue generating behavioral feature information. This process is crucial for ensuring the accuracy of subsequent analysis, because only complete and valid behavioral data can generate accurate behavioral feature information, thus providing a reliable basis for risk assessment. In this way, this invention can effectively filter out invalid or incomplete data, avoiding interference with subsequent security analysis.

[0104] Specifically, behavioral data refers to records of worker behavior acquired from on-site safety monitoring devices. These records can be video data, motion sensor data, or other forms of monitoring data. Integrity detection processing is a data verification process used to check whether the behavioral data contains valid behavioral records. For example, for video behavioral data, integrity detection might involve checking for video continuity, obvious obstructions, or blurriness; for motion sensor data, it might check for the integrity of sensor signals and the presence of data loss. The integrity detection result is a judgment indicator, indicating whether the data meets the conditions for further processing. If the integrity detection result indicates the presence of valid behavioral records, the system will generate behavioral feature information based on this data. Behavioral feature information, obtained through analysis and processing of behavioral data, reflects key characteristics of worker behavior patterns, such as movement frequency, behavioral trajectory, or posture features.

[0105] Preferably, the integrity detection process can be further refined into various specific operations. Taking video behavior data as an example, integrity detection can include action recognition processing, that is, using computer vision technology to identify the actions of people in the video and determine whether a complete action sequence exists. For example, deep learning models (such as convolutional neural networks) can be used to analyze video frames to identify specific actions of workers, such as raising their hands, bending over, or walking. If the model can identify a series of coherent actions, the video data is considered to be complete. For motion sensor data, integrity detection can include time-series analysis of sensor signals to check for missing data or abnormal fluctuations. For example, a threshold can be set; if the sensor data does not change within a certain period of time or the change exceeds the normal range, the data is considered incomplete. Through these specific integrity detection steps, the system can more accurately filter out valid behavioral data, thereby improving the accuracy and reliability of behavioral feature information generation.

[0106] In some embodiments, the behavior monitoring request is a video behavior monitoring request, and the behavior data is behavior video data; and the process of performing integrity detection processing on the behavior data to obtain an integrity detection result includes: performing action recognition processing on the behavior data to obtain an action recognition result as the integrity detection result.

[0107] It should be noted that the behavior monitoring request mentioned in this invention is a video behavior monitoring request, and the behavior data is behavior video data. In this case, integrity detection processing is achieved by performing action recognition processing on the behavior video data. The purpose of action recognition processing is to extract the action features of the workers from the video data, thereby determining whether the video data contains complete and valid action records. This processing method can effectively identify the behavior patterns of workers in the video, providing a basis for subsequent generation of behavior feature information. Using the action recognition result obtained through action recognition processing as the integrity detection result can ensure the quality and usability of the video data, thereby improving the accuracy and reliability of the entire behavior analysis system.

[0108] Specifically, a video behavior monitoring request refers to a request initiated by a target application to conduct video surveillance of the work site and analyze the behavior of the workers. Behavioral video data refers to the behavioral records of workers obtained from video surveillance equipment, usually existing in the form of a video stream. Action recognition processing in integrity detection is a computer vision-based technique used to extract action features from video data. This process typically involves analyzing video frames to identify the movement patterns of people in the video. For example, deep learning algorithms, such as convolutional neural networks (CNNs) or recurrent neural networks (RNNs), can be used to implement action recognition. These algorithms can identify action sequences in the video, such as waving, jumping, running, etc., and determine whether the actions are complete. The integrity detection result is the output of the action recognition processing; it indicates whether the video data contains valid action records. If the video data contains a complete action sequence, the data is considered valid and can be used for further generation of behavioral feature information.

[0109] Preferably, action recognition processing can be implemented using a deep learning model. For example, a model based on a convolutional neural network (CNN) can be constructed, which identifies action features by analyzing pixel information in video frames. The model's input parameters can include the resolution, frame rate, and duration of the video frames. During the model training phase, a large amount of labeled video data can be used as a training set, containing various common worker actions. Through training, the model can learn feature representations of different actions. In practical applications, the collected behavioral video data is input into the trained model, which outputs action recognition results, indicating whether a complete action sequence exists in the video. If a complete action sequence is identified, this result is used as an integrity detection result and further used to generate behavioral feature information. Furthermore, parameters can be set to optimize the accuracy of action recognition, such as setting a confidence threshold for action recognition. Only when the model's confidence in recognizing an action is higher than this threshold is the action considered valid. In this way, the accuracy and reliability of integrity detection can be further improved.

[0110] In some embodiments, generating behavioral feature information based on the behavioral data includes:

[0111] The behavioral data is segmented into frames to obtain a behavioral data sequence;

[0112] For each behavioral data unit in the behavioral data sequence, perform the following steps: determine whether a complete human outline region exists in the behavioral data unit;

[0113] In response to determining that a complete human contour region exists in the behavior data unit, the human contour region in the behavior data unit is cropped to obtain contour data.

[0114] The contour data is color simplified to obtain simplified contour data;

[0115] Brightness detection is performed on the simplified contour data to obtain contour data intensity information;

[0116] Contrast detection is performed on the simplified contour data to obtain contour data comparison information;

[0117] Based on the simplified contour data quality information, the contour data intensity information, and the contour data comparison information, contour data parameters are generated.

[0118] Based on the obtained contour data parameters, select the contour data that meets the preset data parameter conditions from the obtained contour data as the target contour data;

[0119] Based on the simplified contour data quality information, contour data intensity information, and contour data comparison information corresponding to the target contour data, the target contour data is optimized to obtain optimized contour data.

[0120] The optimized contour data is subjected to feature extraction processing to obtain contour data feature information;

[0121] Behavioral feature information is generated based on the contour data feature information and the personnel information.

[0122] It should be noted that the behavioral data processing procedure mentioned in this invention involves a series of complex analyses and optimizations of behavioral video data to generate behavioral feature information. This process includes frame segmentation of the behavioral data, detection and cropping of personnel outline regions, color simplification, brightness and contrast detection, and data optimization and feature extraction based on these detection results. Through these steps, key feature information that accurately reflects the behavioral patterns of workers can be extracted from the raw video data, providing high-quality data support for subsequent behavioral analysis and risk assessment. This method can effectively improve the accuracy and reliability of behavioral analysis while reducing the interference of invalid data on the analysis results.

[0123] Specifically, behavioral data sequences refer to decomposing behavioral video data into a series of individual frame images, each containing behavioral information of the worker at a specific moment. When processing each frame, the first step is to determine if a complete human silhouette region exists, i.e., whether the worker's body outline is complete and clearly visible. If a complete silhouette region exists, it is cropped to obtain the silhouette data. Color simplification involves simplifying the color information in the silhouette data into a more easily processed form, such as converting a color image to grayscale or another limited color range, to reduce data volume and highlight key features. Brightness and contrast detection are image quality assessment operations performed on the simplified silhouette data. Brightness detection evaluates the brightness of the image, while contrast detection evaluates the differences in brightness between different areas of the image. These detection results are used to generate silhouette data parameters, select target silhouette data that meets preset conditions, and perform data optimization processing to ultimately extract silhouette data feature information. This feature information, combined with human information, generates behavioral feature information for subsequent behavioral analysis.

[0124] Preferably, the detection and cropping of person contour regions can be achieved using computer vision techniques. For example, background subtraction or deep learning models (such as object detection models based on convolutional neural networks) can be used to detect person contour regions in each frame of the image. If the detected person contour region is complete and clear, the region is cropped to ensure the accuracy of subsequent processing. Color simplification can be achieved by converting the image to a grayscale image, which can effectively reduce the complexity of color information while preserving the basic structural features of the image. Brightness detection can be achieved by calculating the average brightness value of pixels in the image, and contrast detection can be achieved by calculating the standard deviation of pixel brightness in the image. These detection results will be used to generate contour data parameters; for example, the quality of the contour data can be evaluated by calculating the weighted sum of brightness and contrast values. In the data optimization processing stage, operations such as quality enhancement, intensity adjustment, and contrast enhancement can be performed on the target contour data based on the quality information, intensity information, and contrast information of the contour data. For example, if the brightness of the contour data is low, the visibility of the image can be enhanced by adjusting the brightness value; if the contrast is low, the contour region can be highlighted by adjusting the contrast of the image. Finally, feature extraction is performed on the optimized contour data. Deep learning models (such as convolutional neural networks) can be used to extract feature vectors from the contour data. These feature vectors will serve as the basis for behavioral feature information and will be used for subsequent behavioral analysis and risk assessment.

[0125] In some embodiments, the step of color simplification processing of the contour data to obtain simplified contour data includes:

[0126] The contour data is color simplified to obtain simplified contour data;

[0127] Remove each data point that meets the preset edge conditions from the simplified contour data to obtain the removed contour data;

[0128] For each data point in the simplified contour data, perform the following steps: determine the data point that is right-adjacent to the data point in the simplified contour data as the right-adjacent data point;

[0129] In the simplified contour data, the data points that are adjacent to the data points are defined as the upper adjacent data points;

[0130] Generate a first value and a second value based on the value of the right adjacent data point and the value of the data point;

[0131] Based on the values ​​of the adjacent data points and the values ​​of the data points, a third value and a fourth value are generated;

[0132] The sum of the first, second, third, and fourth values ​​is determined as the fifth value.

[0133] The number of each data point included in the simplified contour data is determined as the total number of data points;

[0134] The ratio of the fifth value to the total number of data points is determined as the contour data quality information corresponding to the simplified contour data.

[0135] It should be noted that the step of color simplification of contour data in this invention further refines the data processing flow. This step not only simplifies the contour data but also optimizes data quality by removing data points that meet preset edge conditions. Furthermore, contour data quality information is generated by calculating the numerical relationships between adjacent data points. This process effectively reduces interference from noise data and, through quantitative evaluation, provides a more accurate basis for subsequent data optimization and feature extraction. In this way, this invention can process behavioral video data more accurately, thereby improving the accuracy and reliability of behavioral feature information generation.

[0136] Specifically, contour data refers to image data containing the contours of workers extracted from behavioral video data. Color simplification processing simplifies complex color information into a more easily processed form, such as converting a color image to a grayscale image or other limited color range. Preset edge conditions refer to certain boundary features defined in the contour data, such as the brightness or color change threshold of edge pixels. Data points that meet these conditions are considered edge data points and may interfere with subsequent processing, so they need to be removed. For each data point, its right neighbor and top neighbor data points refer to the pixels that are horizontally to the right and vertically upward to the current data point in the image matrix. By calculating the relationship between the values ​​of these neighboring data points and the value of the current data point, a series of values ​​can be generated, which reflect the interrelationships between the data points. Finally, by calculating the ratio of the sum of these values ​​to the total number of data points, contour data quality information is obtained, which is used to evaluate the overall quality of the contour data.

[0137] Preferably, color simplification can be achieved by converting a color image to a grayscale image. This method effectively reduces the complexity of color information while preserving the basic structural features of the image. Preset edge conditions can be set to indicate that the brightness change of edge pixels exceeds a certain threshold. For example, when the brightness change exceeds a certain set value, the pixel is considered an edge data point and is removed. When calculating the numerical relationship between adjacent data points, specific calculation rules can be defined. For example, the first value can be the difference between the brightness value of the right adjacent data point and the brightness value of the current data point; the second value can be the ratio of the brightness value of the right adjacent data point to the brightness value of the current data point; the third and fourth values ​​correspond to similar calculations for adjacent data points. By adding these values ​​to obtain a fifth value, and then dividing by the total number of data points, the contour data quality information is obtained. This information can be used for subsequent data optimization processing, such as selecting whether to further enhance the contour data based on the quality information. In this way, the data processing flow can be controlled more precisely, ensuring that the generated behavioral feature information has high quality and high reliability.

[0138] In some embodiments, the step of performing data optimization processing on the target contour data based on the simplified contour data quality information, contour data intensity information, and contour data comparison information corresponding to the target contour data to obtain optimized contour data includes:

[0139] In response to determining that the quality information of the simplified contour data corresponding to the target contour data meets the preset quality optimization conditions, the target contour data is subjected to quality enhancement processing to update the target contour data;

[0140] In response to determining that the contour data intensity information corresponding to the target contour data meets the preset intensity optimization conditions, the target contour data is subjected to intensity adjustment processing to update the target contour data;

[0141] In response to determining that the contour data comparison information corresponding to the target contour data meets the preset comparison optimization conditions, the target contour data is subjected to comparison enhancement processing to update the target contour data;

[0142] The updated target contour data is then normalized in size to update the target contour data.

[0143] The updated target contour data is used as the optimized contour data.

[0144] It should be noted that the data optimization steps for the target contour data in this invention aim to further improve the quality and accuracy of behavioral feature information. By performing targeted optimization based on the quality, intensity, and contrast information of the contour data, such as quality enhancement, intensity adjustment, and contrast enhancement, it can be ensured that the optimized contour data more clearly and accurately reflects the behavioral characteristics of the workers. Furthermore, size normalization processing of the updated target contour data unifies the data format, making it more suitable for subsequent feature extraction and analysis processes. This series of optimization operations provides a strong guarantee for generating high-quality behavioral feature information, thereby improving the reliability and effectiveness of the entire behavioral analysis system.

[0145] Specifically, target contour data refers to contour data that meets preset data parameter conditions after preliminary processing (such as color simplification and noise removal). Contour data quality information reflects the overall quality of the contour data, such as image sharpness, integrity, and noise level. Contour data intensity information usually refers to the average level of image brightness or grayscale values, reflecting the overall brightness and darkness of the image. Contour data contrast information indicates the degree of brightness difference between different areas in the image; higher contrast helps to highlight contour features. Preset quality optimization conditions, preset intensity optimization conditions, and preset contrast optimization conditions are thresholds or standards set according to actual application needs to determine whether corresponding optimization processing of the target contour data is required. For example, if the quality information of the target contour data is lower than a certain set threshold, quality enhancement processing is triggered; if the intensity information is not within a reasonable range, intensity adjustment processing is performed; if the contrast information is insufficient, contrast enhancement processing is performed. Size normalization processing refers to adjusting the size of the target contour data to a uniform standard size for subsequent processing and analysis.

[0146] Preferably, the specific steps of data optimization processing can be further refined. For example, in quality enhancement processing, image filtering techniques, such as median filtering or Gaussian filtering, can be used to reduce noise and improve image quality. Specific parameter settings include the size and shape of the filter; for example, the window size of the median filter can be adjusted according to the resolution and noise level of the contour data. In intensity adjustment processing, histogram equalization techniques can be used to adjust the brightness distribution of the image, making the intensity information of the image more uniform. Contrast enhancement processing can be achieved by adjusting the contrast range of the image, for example, by enhancing the contrast of the image through linear stretching or nonlinear transformation (such as gamma correction). Specific parameter settings include the range of contrast adjustment and the parameters of the transformation function. In size normalization processing, the width and height of the target contour data can be adjusted to a preset standard size, for example, uniformly adjusting all contour data to a 128×128 pixel image. This process can be implemented through image interpolation algorithms, with specific parameters including the interpolation method (such as nearest neighbor interpolation, bilinear interpolation, etc.) and the target size. Through these specific optimization processing steps, the quality and consistency of the target contour data can be ensured, thereby providing high-quality data support for subsequent behavioral feature extraction and analysis.

[0147] In some embodiments, the behavior monitoring request is an action behavior monitoring request, and the behavior data is motion sensor data; and the step of performing integrity detection processing on the behavior data to obtain an integrity detection result includes: performing action feature extraction processing on the behavior data to obtain action feature results as integrity detection results; and the step of generating behavior feature information based on the behavior data includes:

[0148] The behavioral data is subjected to noise filtering and invalid data removal to obtain processed data;

[0149] The processed data is subjected to feature extraction to obtain an action feature vector;

[0150] For each feature dimension in the action feature vector, perform the following steps: determine the feature value corresponding to the feature dimension;

[0151] Determine the feature weights corresponding to the feature dimensions;

[0152] A comprehensive feature score is generated based on the feature values ​​and feature weights;

[0153] Select the comprehensive feature score that meets the preset scoring conditions from the various generated comprehensive feature scores as the target comprehensive feature score;

[0154] The target comprehensive feature score and the personnel information are determined as behavioral feature information.

[0155] It should be noted that the behavioral data processing method for motion monitoring requests in this invention provides a method for integrity detection and behavioral feature information generation based on motion sensor data. Motion sensor data typically contains worker motion information, such as acceleration and angular velocity. By performing noise filtering, invalid data removal, and feature extraction on this data, behavioral feature information reflecting the worker's behavioral patterns can be generated. This method is particularly suitable for scenarios where behavioral data cannot be obtained through video surveillance, such as in low-light environments or work sites requiring privacy protection. Through comprehensive feature scoring, the worker's behavioral characteristics can be assessed more accurately, providing a reliable basis for subsequent risk assessment.

[0156] Specifically, a motion monitoring request refers to a request initiated by the target application to monitor the actions of workers at the work site. Motion sensor data refers to the motion information of workers collected by motion sensors (such as accelerometers and gyroscopes). This data is typically recorded in time series format, including the amplitude, frequency, and direction of the workers' movements. Integrity detection processing involves analyzing the motion sensor data to determine whether it contains complete and valid motion features. Noise filtering and invalid data removal processes aim to remove interference signals and unreasonable data points from the data, thereby improving data quality. Feature extraction processing transforms the processed motion data into feature vectors, where each feature dimension corresponds to a specific feature of the motion data. Feature weights refer to the proportion of importance of each feature dimension in the behavioral feature information and can be set according to the actual application scenario. A comprehensive feature score is a quantitative indicator used to evaluate the behavioral characteristics of workers, calculated by combining feature values ​​and feature weights. Preset scoring conditions refer to the standards set according to actual needs for filtering behavioral feature information, such as setting a scoring threshold; only comprehensive feature scores that meet this threshold will be selected as the target comprehensive feature score.

[0157] Preferably, the processing of motion sensor data can be further refined. For example, in the noise filtering and invalid data removal stage, digital filtering techniques can be used, such as low-pass filters to remove high-frequency noise, or data points exceeding the normal range can be removed by setting a threshold. In the feature extraction stage, machine learning algorithms, such as Principal Component Analysis (PCA) or Support Vector Machine (SVM), can be used to extract key features of the motion data. Specific parameter settings include the filter cutoff frequency and the number of principal components in the PCA. The feature weights can be adjusted according to the physical meaning of the motion data and the actual application scenario; for example, the weight of acceleration data may be higher than that of angular velocity data. The comprehensive feature score can be calculated by multiplying the feature value of each feature dimension by its corresponding feature weight and then summing the results. The selection of the target comprehensive feature score can be achieved by setting a scoring threshold; for example, features with scores higher than a certain fixed value can be selected as the target comprehensive feature score. Through these specific processing steps, behavioral feature information can be generated more accurately, thereby improving the accuracy and reliability of behavioral analysis.

[0158] The various embodiments of the present invention have the following beneficial effects: The present invention can collect and intelligently analyze the behavioral data of workers in real time, and significantly improve the accuracy of on-site safety monitoring by dynamically generating behavioral feature information and risk assessment mechanisms. When abnormal behavior is detected, verification rules can be automatically generated and matched with historical data to quickly identify potential risks, thereby triggering an early warning mechanism in a timely manner and effectively reducing the possibility of safety accidents.

[0159] Furthermore, this invention can adaptively process data for different monitoring needs (such as video behavior or sensor motion data), ensuring the reliability of behavior analysis through steps such as data optimization, feature extraction, and similarity matching. Simultaneously, it can combine personnel location information with work area conditions to achieve precise monitoring, and improve data effectiveness through techniques such as quality enhancement and intensity adjustment, further enhancing the timeliness and accuracy of safety warnings.

[0160] The electronic devices in some embodiments of the present invention may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers.

[0161] Electronic devices may include processing units (such as central processing units, graphics processing units, etc.) that can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) or loaded from storage devices into random access memory (RAM). RAM also stores various programs and data required for the operation of the electronic device. The processing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0162] Typically, the following devices can be connected to the I / O interface: input devices such as touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices such as liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices such as magnetic tapes, hard drives, etc.; and communication devices. Communication devices allow electronic devices to communicate wirelessly or wiredly with other devices to exchange data.

[0163] Furthermore, the storage medium in the embodiments of this application stores program instructions capable of implementing all the above methods. These program instructions can be stored in the storage medium in the form of a software product, including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, or terminal devices such as computers, servers, mobile phones, and tablets.

[0164] The above description is merely a selection of preferred embodiments of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention as described in the embodiments is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. A workplace safety analysis method based on personnel behavior analysis, characterized in that, include: In response to receiving a behavior monitoring request through a target application associated with a work site safety monitoring device, the behavior monitoring request is parsed to obtain the worker information corresponding to the behavior monitoring request, wherein the behavior monitoring request is a request to perform behavior monitoring on the current work site; Obtain the behavioral data of the workers; Based on the behavioral data, behavioral feature information is generated; An analysis pop-up window corresponding to the behavioral characteristic information is displayed, wherein the behavioral characteristic information and risk assessment configuration controls are displayed in the analysis pop-up window; Based on the risk level information corresponding to the configured risk assessment control, the behavioral characteristic information is assessed for risk. In response to an abnormal behavior request detected by the on-site safety monitoring device, a behavior verification rule is generated based on a preset behavior library; The behavior verification rules are sent to the work site safety monitoring device, so that the work site safety monitoring device displays the behavior verification rules on an external display. The system receives the behavior data to be verified, which corresponds to the behavior verification rules, sent by the on-site safety monitoring device as the information of the person to be identified. The behavior data to be verified is the behavior record of the worker during the monitoring period. In response to obtaining the information of the person to be identified through the work site safety monitoring device, the information of the person to be identified is matched according to the stored behavioral feature information to obtain a matching result. The matching process of the information of the person to be identified according to the stored behavioral feature information to obtain a matching result includes: performing noise filtering and invalid data removal on the behavioral data to be verified to obtain processed behavioral data. The processed behavioral data is subjected to feature extraction to obtain a behavioral feature vector; Generate the behavioral similarity between the behavioral feature vector and the historical feature vectors included in each behavioral feature information, and obtain a similarity set; The similarity scores in the similarity set that satisfy a first preset similarity condition are determined as the target similarity scores, wherein the first preset similarity condition is that the similarity score is the maximum value in the similarity set; In response to determining that the target similarity satisfies a second preset similarity condition, the feature vector corresponding to the target similarity is determined as a matching feature vector, wherein the second preset similarity condition is that the target similarity is greater than a preset similarity threshold; Based on the personnel information corresponding to the matching feature vector, a matching result is generated; In response to determining that the matching result indicates a successful match, the early warning component of the on-site safety monitoring device is controlled to perform a safety early warning operation.

2. The method according to claim 1, characterized in that, The operator information includes operator identification and location information; and before acquiring the operator's behavior data, the method further includes: In response to detecting a selection operation of the launch control on the behavior monitoring page, the location information of the operator corresponding to the behavior monitoring request is obtained according to the operator identifier, wherein the behavior monitoring page corresponds to the behavior monitoring request, and the operator information, including the operator's name, is displayed on the behavior monitoring page; Determine whether the location information meets the preset range conditions corresponding to the target work area; In response to determining that the location information meets the preset range condition, a behavior data collection instruction corresponding to the behavior monitoring request is sent to the work site safety monitoring device; and the acquisition of the behavior data of the worker includes: in response to receiving feedback information indicating consent to collection corresponding to the behavior data collection instruction, acquiring the behavior data of the worker.

3. The method according to claim 2, characterized in that, The step of generating behavioral feature information based on the behavioral data includes: The behavioral data is subjected to integrity detection processing to obtain integrity detection results; In response to determining that the integrity detection result indicates the existence of a valid behavior record, behavior feature information is generated based on the behavior data.

4. The method according to claim 3, characterized in that, The behavior monitoring request is a video behavior monitoring request, and the behavior data is behavior video data; The process of performing integrity detection on the behavioral data to obtain an integrity detection result includes: performing action recognition processing on the behavioral data to obtain an action recognition result as the integrity detection result.

5. The method according to claim 4, characterized in that, The step of generating behavioral feature information based on the behavioral data includes: The behavioral data is segmented into frames to obtain a behavioral data sequence; For each behavioral data unit in the behavioral data sequence, perform the following steps: determine whether a complete human outline region exists in the behavioral data unit; In response to determining that a complete human contour region exists in the behavior data unit, the human contour region in the behavior data unit is cropped to obtain contour data. The contour data is color simplified to obtain simplified contour data; Brightness detection is performed on the simplified contour data to obtain contour data intensity information; Contrast detection is performed on the simplified contour data to obtain contour data comparison information; Based on the simplified contour data quality information, the contour data intensity information, and the contour data comparison information, contour data parameters are generated. Based on the obtained contour data parameters, select the contour data that meets the preset data parameter conditions from the obtained contour data as the target contour data; Based on the simplified contour data quality information, contour data intensity information, and contour data comparison information corresponding to the target contour data, the target contour data is optimized to obtain optimized contour data. The optimized contour data is subjected to feature extraction processing to obtain contour data feature information; Behavioral feature information is generated based on the contour data feature information and the personnel information.

6. The method according to claim 5, characterized in that, The step of performing color simplification processing on the contour data to obtain simplified contour data includes: The contour data is color simplified to obtain simplified contour data; Remove each data point that meets the preset edge conditions from the simplified contour data to obtain the removed contour data; For each data point in the simplified contour data, perform the following steps: determine the data point that is right-adjacent to the data point in the simplified contour data as the right-adjacent data point; In the simplified contour data, the data points that are adjacent to the data points are defined as the upper adjacent data points; Generate a first value and a second value based on the value of the right adjacent data point and the value of the data point; Based on the values ​​of the adjacent data points and the values ​​of the data points, a third value and a fourth value are generated; The sum of the first, second, third, and fourth values ​​is determined as the fifth value. The number of each data point included in the simplified contour data is determined as the total number of data points; The ratio of the fifth value to the total number of data points is determined as the contour data quality information corresponding to the simplified contour data.

7. The method according to claim 5, characterized in that, The step of optimizing the target contour data based on the simplified contour data quality information, contour data intensity information, and contour data comparison information corresponding to the target contour data to obtain optimized contour data includes: In response to determining that the quality information of the simplified contour data corresponding to the target contour data meets the preset quality optimization conditions, the target contour data is subjected to quality enhancement processing to update the target contour data; In response to determining that the contour data intensity information corresponding to the target contour data meets the preset intensity optimization conditions, the target contour data is subjected to intensity adjustment processing to update the target contour data; In response to determining that the contour data comparison information corresponding to the target contour data meets the preset comparison optimization conditions, the target contour data is subjected to comparison enhancement processing to update the target contour data; The updated target contour data is then normalized in size to update the target contour data. The updated target contour data is used as the optimized contour data.

8. The method according to claim 3, characterized in that, The behavior monitoring request is an action behavior monitoring request, and the behavior data is motion sensor data; The process of performing integrity detection on the behavioral data to obtain an integrity detection result includes: performing action feature extraction on the behavioral data to obtain action feature results as integrity detection results; and generating behavioral feature information based on the behavioral data includes: The behavioral data is subjected to noise filtering and invalid data removal to obtain processed data; The processed data is subjected to feature extraction to obtain an action feature vector; For each feature dimension in the action feature vector, perform the following steps: determine the feature value corresponding to the feature dimension; Determine the feature weights corresponding to the feature dimensions; A comprehensive feature score is generated based on the feature values ​​and feature weights; Select the comprehensive feature score that meets the preset scoring conditions from the various generated comprehensive feature scores as the target comprehensive feature score; The target comprehensive feature score and the personnel information are determined as behavioral feature information.

9. An electronic device, comprising: One or more processors; Storage device, on which one or more programs are stored, When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-8.

10. A computer-readable medium having a computer program stored thereon, wherein, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-8.

Images