Dangerous behavior supervision method and equipment based on electronic fence, and medium

Abstract

The embodiment of the invention discloses a dangerous behavior supervision method and equipment based on an electronic fence, and a medium, belongs to the technical field of safety management, and solves the problems that fixed electronic fence division is difficult to adapt to storage changes, so that part of regions are overly supervised, resources are wasted, and part of high-risk regions are insufficient in supervision, and security holes exist. Comprising the steps of dividing a storage area into a plurality of supervision areas based on layout information and goods information of the storage area, and performing electronic fence division based on danger levels corresponding to the plurality of supervision areas. And obtaining corresponding monitoring information in each electronic fence, and determining dangerous behavior severity levels corresponding to the plurality of supervision areas based on the monitoring information. And under the condition that the severity level of the dangerous behavior does not meet the preset condition, performing supervision level division on the supervision area based on the occurrence position of the dangerous behavior. And based on the supervision level, performing dynamic regulation and control on the electronic fence of the supervision area so as to perform hierarchical supervision on the supervision area.

Description

Technical Field

[0001] The present application relates to the field of safety management technology, and in particular to a dangerous behavior supervision method, device and medium based on electronic fences. Background Art

[0002] With the rapid development of the logistics industry, warehousing, as a key link in the supply chain, has become increasingly large and complex. This is especially true for the warehousing of special items, such as flammable, explosive, toxic and hazardous chemicals, and the storage and management of valuables, which places extremely high demands on safety and regulatory accuracy.

[0003] Currently, warehouse safety supervision primarily relies on a combination of traditional physical fencing and manual inspections, or electronic fencing systems with fixed area divisions. However, these approaches have significant drawbacks. In traditional physical fencing and manual inspection models, physical fencing only provides regional isolation, making it difficult to proactively monitor and provide early warnings for dangerous behavior. Furthermore, manual inspections suffer from long intervals, high subjectivity, and low efficiency, making it difficult to detect dangerous behavior in real time in complex warehouse operating environments.

[0004] While existing electronic fence systems have improved regulatory efficiency to a certain extent, in real-world warehousing scenarios, the layout of storage areas, the types and locations of goods, and the patterns of human and vehicle activity are all subject to dynamic change. For example, as new goods arrive and the storage layout adjusts, the risk levels of different areas also change. However, fixed electronic fence systems struggle to adapt to these changes, resulting in over-regulation in some areas and wasted resources, while under-regulation in high-risk areas creates security vulnerabilities. Summary of the Invention

[0005] The embodiments of the present application provide a dangerous behavior supervision method, device and medium based on electronic fences, which are used to solve the following technical problems: fixed electronic fence divisions are difficult to adapt to changes in the storage environment, resulting in excessive supervision of some areas and waste of resources, while some high-risk areas are under-supervised and have security loopholes.

[0006] The embodiments of this application adopt the following technical solutions:

[0007] The embodiment of the present application provides a dangerous behavior supervision method based on electronic fences. It includes dividing the storage area into multiple supervision areas based on the layout information and product information of the storage area, and dividing the electronic fences based on the danger levels corresponding to the multiple supervision areas. Obtaining the monitoring information corresponding to each electronic fence, and determining the severity level of dangerous behaviors corresponding to the multiple supervision areas based on the monitoring information. When the severity level of the dangerous behavior does not meet the preset level conditions, the supervision area is divided into supervision levels based on the location where the dangerous behavior occurs. Based on the supervision level, the electronic fences of the supervision area are dynamically adjusted to perform hierarchical supervision of the supervision area.

[0008] The embodiment of the present application divides the supervision area by the layout information and product information of the storage area, and then divides the electronic fence according to the danger level, which can accurately match the supervision needs of different areas. By obtaining the monitoring information within the electronic fence to determine the severity level of dangerous behavior, the safety dynamics of the warehouse can be grasped in real time. In addition, the supervision level of the supervision area is divided based on the location where the dangerous behavior occurs, and the electronic fence can be dynamically adjusted to form a hierarchical supervision system. Differentiated protection measures are taken for areas with different supervision levels. While reducing supervision resources, the overall safety protection capabilities of the warehouse are enhanced and the probability of safety accidents is reduced.

[0009] In one implementation of the present application, the storage area is divided into multiple supervision areas based on the layout information and product information of the storage area, and electronic fence division is performed based on the danger levels corresponding to the multiple supervision areas, specifically including: obtaining a first supervision area distribution based on the layout information of the storage area; wherein the first supervision area distribution includes the functional area type, functional area location, functional area area and functional area shape; based on the product information, the stored goods are classified, and the second supervision area distribution corresponding to the same type of goods is determined in the goods storage area; wherein the second supervision area distribution includes the stored goods type, storage location and storage area; according to the first supervision area distribution and the first distribution weight, the first danger level corresponding to the first supervision area distribution is obtained; according to the second supervision area distribution and the second distribution weight, the second danger level corresponding to the second supervision area distribution is obtained; based on the first danger level and the second danger level, the electronic fence supervision level corresponding to each supervision area is determined, and electronic fence division is performed.

[0010] In one implementation of the present application, the severity levels of dangerous behaviors corresponding to multiple regulatory areas are determined based on monitoring information, specifically including: extracting dangerous features from the monitoring information in real time through a preset deep learning model; wherein the dangerous features include dangerous behavior types and the number of dangerous behavior occurrences corresponding to each dangerous behavior type; matching the extracted dangerous features with time information and spatial information to construct a spatiotemporal dangerous feature data set; wherein the spatiotemporal dangerous feature data set includes at least the timestamp of the danger occurrence and the location of the danger occurrence; based on the spatiotemporal dangerous feature data set, determining the corresponding danger level weights in a preset danger level setting table; and determining the danger level severity levels corresponding to multiple regulatory areas according to the spatiotemporal dangerous feature data set and the danger level weights.

[0011] In one implementation of the present application, after extracting dangerous features from monitoring information in real time through a preset deep learning model, the method also includes: when the dangerous feature falls within a preset alarm feature range, determining a bounding box of the feature target in the video frame of the monitoring information; obtaining a target image through the bounding box of the feature target, tracking the behavior of the feature target in the target image, and obtaining a motion trajectory of the feature target; obtaining a predicted motion trend based on the motion trajectory of the feature target; and determining the dangerous consequence level corresponding to the dangerous feature based on the predicted motion trend, so as to control the electronic fence to alarm according to the dangerous consequence level.

[0012] In one implementation of the present application, the severity levels of dangerous behaviors corresponding to multiple regulatory areas are determined based on the spatiotemporal hazard characteristic data set and the hazard degree weight, specifically including: determining the spatiotemporal hazard characteristic data corresponding to each regulatory area; assigning type weights based on the dangerous behavior types corresponding to the spatiotemporal hazard characteristic data; and assigning number weights based on the number of dangerous behaviors corresponding to each dangerous behavior type; obtaining the severity level of dangerous behaviors corresponding to the regulatory area based on the spatiotemporal hazard characteristic data, type weights and number weights.

[0013] In one implementation of the present application, when the severity level of the dangerous behavior does not meet the preset level conditions, the supervision area is divided into supervision levels based on the location where the dangerous behavior occurs, specifically including: marking the location where the dangerous behavior occurs in the electronic map corresponding to the storage area; determining the dangerous impact range based on the dangerous behavior type corresponding to the dangerous behavior; drawing the dangerous impact range corresponding to each dangerous behavior in the supervision area with the marked point as the origin; and superimposing each dangerous impact range to divide the supervision area into supervision levels based on the superposition result.

[0014] In one implementation of the present application, the electronic fence of the supervision area is dynamically adjusted based on the supervision level to carry out hierarchical supervision of the supervision area, specifically including: determining the dangerous behavior types corresponding to areas with different supervision levels; wherein the dangerous behavior types include at least operational hazards and environmental hazards; and constructing an intelligent agent group based on the dangerous behavior types; wherein the intelligent agent group is used to adjust the electronic fence configuration of multiple different danger dimensions; based on the number of occurrences corresponding to each dangerous behavior type in areas with different supervision levels, different adjustment values ​​are assigned to the intelligent agent group to dynamically adjust the electronic fence configuration based on the adjustment value.

[0015] In one implementation of the present application, hierarchical supervision is performed on the supervision area, specifically including: when a dangerous behavior alarm occurs in the supervision area, the associated supervision area corresponding to the supervision area is determined based on the spatial position of the electronic fence; according to the characteristics of the dangerous behavior, the electronic fence of the associated supervision area is configured and adjusted.

[0016] An embodiment of the present application provides a dangerous behavior supervision device based on an electronic fence, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can: divide the storage area into multiple supervision areas based on the layout information and product information of the storage area, and divide the electronic fences based on the danger levels corresponding to the multiple supervision areas. Obtain the monitoring information corresponding to each electronic fence, and determine the severity level of the dangerous behavior corresponding to the multiple supervision areas based on the monitoring information. In the case where the severity level of the dangerous behavior does not meet the preset level conditions, the supervision area is divided into supervision levels based on the location where the dangerous behavior occurs. Based on the supervision level, the electronic fences of the supervision area are dynamically adjusted to perform hierarchical supervision of the supervision area.

[0017] The embodiment of the present application provides a non-volatile computer storage medium storing computer executable instructions, wherein the computer executable instructions are configured to: divide the storage area into multiple supervision areas based on the layout information and product information of the storage area, and perform electronic fence division based on the danger levels corresponding to the multiple supervision areas. Obtain monitoring information corresponding to each electronic fence, and determine the severity level of dangerous behaviors corresponding to the multiple supervision areas based on the monitoring information. When the severity level of the dangerous behavior does not meet the preset level conditions, the supervision area is divided into supervision levels based on the location where the dangerous behavior occurs. Based on the supervision level, the electronic fences of the supervision area are dynamically adjusted to perform hierarchical supervision of the supervision area.

[0018] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects: The embodiments of the present application divide the supervision area by the layout information and product information of the storage area, and then divide the electronic fence according to the danger level, which can accurately match the supervision needs of different areas. By obtaining the monitoring information within the electronic fence to determine the severity level of the dangerous behavior, the safety dynamics of the warehouse can be grasped in real time. In addition, the supervision level of the supervision area is divided based on the location where the dangerous behavior occurs, and the electronic fence can be dynamically adjusted to form a hierarchical supervision system. Differentiated protection measures are taken for areas with different supervision levels. While reducing supervision resources, the overall safety protection capabilities of the warehouse are enhanced and the probability of safety accidents is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments described in the present application. For those skilled in the art, other drawings can be obtained based on these drawings without inventive work. In the drawings:

[0020] Figure 1 A flowchart of a dangerous behavior supervision method based on an electronic fence provided in an embodiment of the present application;

[0021] Figure 2 A schematic structural diagram of a dangerous behavior monitoring device based on an electronic fence provided in an embodiment of the present application.

[0022] Reference numerals:

[0023] 200: Dangerous behavior monitoring device based on electronic fence, 201: Processor, 202: Memory. DETAILED DESCRIPTION

[0024] The embodiments of the present application provide a method, device, and medium for monitoring dangerous behaviors based on electronic fences.

[0025] In order to enable those skilled in the art to better understand the technical solutions in this application, the following will clearly and completely describe the technical solutions in the embodiments of this application in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0026] The technical solutions proposed in the embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0027] Figure 1A flowchart of a dangerous behavior supervision method based on an electronic fence is provided in an embodiment of the present application. Figure 1 As shown, the dangerous behavior supervision method based on electronic fence includes the following steps:

[0028] S101. Based on the layout information and product information of the storage area, the storage area is divided into multiple supervision areas, and electronic fences are performed based on the danger levels corresponding to the multiple supervision areas.

[0029] In one embodiment of the present application, a first supervision area distribution is obtained based on the layout information of the storage area; wherein the first supervision area distribution includes the functional area type, functional area location, functional area area and functional area shape. Based on the goods information, the stored goods are classified, and a second supervision area distribution corresponding to the same type of goods is determined in the goods storage area; wherein the second supervision area distribution includes the stored goods type, storage location and storage area. According to the first supervision area distribution and the first distribution weight, a first danger level corresponding to the first supervision area distribution is obtained. According to the second supervision area distribution and the second distribution weight, a second danger level corresponding to the second supervision area distribution is obtained. Based on the first danger level and the second danger level, the electronic fence supervision level corresponding to each supervision area is determined.

[0030] Specifically, the original layout data of the storage area is obtained, including the location information of walls, passages, doors and windows, fire-fighting facilities, etc. According to the original layout data, the area except the goods storage area in the storage area is divided into multiple different functional areas, such as path channel area, loading and unloading area, rest area, tool placement area, etc., to obtain the first supervision area distribution.

[0031] Furthermore, in the goods storage area of ​​the storage area, the real-time storage location of the goods is obtained through RFID, barcode or manual entry, the storage areas of the same type of goods are clustered and analyzed, and adjacent and continuous storage points are merged to form a second supervision area distribution.

[0032] In this embodiment, each functional area has an inherent risk value associated with it: for example, the risk value of the path passage area is greater than that of the loading and unloading area. Furthermore, based on the distance of each first supervision area from the fire source, entrances and exits, and crowded areas, different weights are assigned to different functional areas in the first supervision area to construct a first distribution weight. Based on the first weight distribution and the risk value corresponding to each functional area, the first hazard level for each functional area in the first supervision area is determined.

[0033] Furthermore, in the embodiments of the present application, each product storage area is assigned an inherent risk value. For example, the risk value of a chemical storage area is greater than that of an ordinary storage area. Furthermore, based on the distance of each second supervision area from fire sources, entrances and exits, and crowded areas, different weights are assigned to different product storage areas in the second supervision area to construct a second distribution weight. Based on the second weight distribution and the risk value corresponding to each storage area, a second hazard level for each storage area in the second supervision area is determined.

[0034] Furthermore, based on the first and second danger levels, the corresponding electronic fence supervision levels of each supervision area are determined to perform the initial electronic fence division of the storage area. The higher the danger level, the higher the electronic fence supervision level.

[0035] S102: Obtain monitoring information corresponding to each electronic fence, and determine the severity levels of dangerous behaviors corresponding to multiple supervision areas based on the monitoring information.

[0036] In one embodiment of the present application, a preset deep learning model is used to extract danger features from monitoring information in real time, wherein the danger features include the types of dangerous behaviors and the number of dangerous behaviors corresponding to each dangerous behavior type. The extracted danger features are matched with time information and spatial information to construct a spatiotemporal danger feature dataset; wherein the spatiotemporal danger feature dataset includes at least the timestamp of the danger occurrence and the location of the danger occurrence. Based on the spatiotemporal danger feature dataset, the corresponding danger level weight is determined in the preset danger level setting table. According to the spatiotemporal danger feature dataset and the danger level weight, the severity levels of dangerous behaviors corresponding to multiple regulatory areas are determined.

[0037] Specifically, real-time video streams are acquired from various surveillance cameras, along with real-time data from various sensors, such as temperature and humidity sensors, smoke sensors, and pressure sensors. Pre-configured deep learning models are used to extract real-time hazard features from this monitoring information. For example, convolutional neural networks can be used to analyze video images and identify hazard features, such as unusual human behavior, the presence of fire sources, abnormal cargo conditions, and abnormal parameter changes indicated by sensor data. The system then accumulates the number of instances of the same type of dangerous behavior.

[0038] Furthermore, extracted hazard signatures are timestamped to record the exact moment they occurred. Hazard signatures are then sorted chronologically. Hazard signatures are combined with the spatial information of the storage area to determine the specific location of the hazard. Using a map and coordinate system of the storage area, hazard signatures are accurately mapped to their corresponding spatial locations. Correlation analysis is performed between temporal and spatial information to determine the propagation and evolution of hazard signatures across time and space. By integrating hazard signatures, temporal information, and spatial information, a spatiotemporal hazard signature dataset is constructed. This dataset contains fields such as the timestamp of the hazard occurrence, location information, and a description of the hazard signature.

[0039] Furthermore, the embodiment of the present application is provided with a preset danger level setting table, which covers different types of danger features and their corresponding danger level levels. For example, different danger level weights are set for danger features such as fire hazards and chemical leaks. The danger features in the spatiotemporal danger feature dataset are matched with the preset danger level setting table, and the corresponding danger level weight is searched in the setting table according to the severity of the danger feature. Based on the spatiotemporal danger feature dataset and the danger level weight, the severity level of dangerous behaviors corresponding to multiple regulatory areas is determined.

[0040] In one embodiment of the present application, spatiotemporal hazard signature data corresponding to each regulatory area is determined. Based on the type of dangerous behavior corresponding to the spatiotemporal hazard signature data, a type weight is assigned, and based on the number of dangerous behavior occurrences corresponding to each dangerous behavior type, a frequency weight is assigned. Based on the spatiotemporal hazard signature data, the type weight, and the frequency weight, the severity level of the dangerous behavior corresponding to the regulatory area is determined.

[0041] Specifically, based on this spatiotemporal hazard signature data, we extract records of dangerous behaviors corresponding to the regulatory areas, parse the timestamp information, and convert the location coordinates. Based on the type of dangerous behavior corresponding to each regulatory area, we assign a type weight to each regulatory area. Each type has a corresponding type weight that reflects its inherent risk level. For example, the type weight for fireworks will be higher than that for ordinary illegal operations. Secondly, based on the number of occurrences of each type of dangerous behavior within the regulatory area, we assign a number weight to each dangerous behavior, where a higher number of occurrences corresponds to a higher number weight.

[0042] Furthermore, based on the inherent risk value corresponding to each dangerous behavior, the assigned type weight and the number weight, the risk value corresponding to each dangerous behavior in the supervision area is determined, and the risk values ​​corresponding to each dangerous behavior in the supervision area are summed up to obtain the total risk value corresponding to the supervision area. The total risk value is compared with the risk level range to determine the severity level of the dangerous behavior corresponding to the supervision area.

[0043] In one embodiment of the present application, after real-time hazard feature extraction is performed on monitoring information, if the hazard feature falls within a preset alarm feature range, a bounding box of the characteristic target is determined in the video frame of the monitoring information. A target image is obtained using the bounding box of the characteristic target, and the behavior of the characteristic target is tracked in the target image to obtain a motion trajectory of the characteristic target. Based on the motion trajectory of the characteristic target, a predicted motion trend is obtained. Based on the predicted motion trend, the hazard consequence level corresponding to the hazard feature is determined, and an electronic fence is controlled to issue an alarm based on the hazard consequence level.

[0044] Specifically, after extracting dangerous features, the deep learning model will classify and predict the identified targets. Based on the large amount of image data features learned during pre-training, the model identifies the target and determines its category, such as dangerous features of human operation or fire source. If the identified dangerous features fall within the preset alarm feature range, meaning the current dangerous features present a significant hidden danger, the deep learning model processes the video frames of the monitoring information. The model uses a convolutional neural network to extract features from the image, determine the bounding box position of the feature target in the video frame, and represent the bounding box range with the upper left and lower right coordinates (x1, y1, x2, y2), separating the feature target from the background.

[0045] Furthermore, a deep learning-based target tracking algorithm is used to track the target's behavior within the image. Feature similarity is calculated to match the same target in consecutive video frames. A Kalman filter algorithm is then used to predict the target's position in the next frame. The target's motion parameters are continuously updated to obtain the target's trajectory within the video sequence, recording the target's position coordinate changes at different time points.

[0046] Furthermore, a deep learning model is trained on motion trajectory data. The model learns the time series patterns and changes in historical trajectory data and, by analyzing the current trajectory segment, predicts the future motion trend of the characteristic target. Based on this future motion trend, the scope of the dangerous behavior and the resulting losses are determined. Based on the scope of the impact and the resulting losses, the dangerous consequence level is determined, and different levels of warnings are issued for dangerous behaviors with different dangerous consequence levels.

[0047] When the level of dangerous consequences is low, the electronic fence will issue a low-intensity alarm to remind relevant personnel to pay attention; when the level of dangerous consequences is high, the electronic fence will immediately issue a high-intensity alarm, link other safety equipment, and push the alarm information to management personnel so that timely measures can be taken to deal with potential dangers.

[0048] S103. When the severity level of the dangerous behavior does not meet the preset level conditions, the supervision level of the supervision area is divided based on the location where the dangerous behavior occurs.

[0049] In one embodiment of the present application, the locations where dangerous behaviors occur are marked on an electronic map corresponding to a storage area. Based on the dangerous behavior type corresponding to the dangerous behavior, a dangerous impact range is determined. With the marked point as the origin, the dangerous impact range corresponding to each dangerous behavior is drawn within the supervision area. The dangerous impact ranges are then regionally superimposed, and the supervision area is classified into a supervision level based on the superposition results.

[0050] Specifically, if the severity level of a dangerous behavior exceeds a preset level, the location of the dangerous behavior captured by the monitoring system is mapped to the coordinate system of the warehouse electronic map. Based on the type of dangerous behavior, the corresponding dangerous impact range is determined in the preset impact range table and the impact range of each dangerous behavior is marked on the electronic map. For example, for fireworks, the impact range is circular with a radius of 20 meters; for cargo collapse, the impact range is fan-shaped with a radius of 10 meters and an angle of 120 degrees.

[0051] Furthermore, the electronic map is divided into grid units of a preset size, such as 5m×5m grid units, and each unit grid is divided into a distance supervision level based on the distance from the marked point, wherein the closer the distance to the marked point, the higher the corresponding supervision level. Secondly, based on the impact range corresponding to each dangerous behavior, the superposition area is determined, and based on the number of superpositions corresponding to each grid unit, the superposition supervision level corresponding to each unit grid is determined. The distance supervision level corresponding to each grid unit is compared with the superposition supervision level, and the highest supervision level is used as the supervision level of the grid unit. In this way, the supervision area is divided into multiple grid units, each grid unit corresponds to a supervision level, and different levels of supervision are performed on different grid units in the same supervision area according to different supervision levels.

[0052] S104. Based on the supervision level, dynamically adjust the electronic fence of the supervision area to perform hierarchical supervision of the supervision area.

[0053] In one embodiment of the present application, dangerous behavior types corresponding to areas with different regulatory levels are determined; the dangerous behavior types include at least operational hazards and environmental hazards. Furthermore, an intelligent agent group is constructed based on the dangerous behavior types; the intelligent agent group is used to adjust the configuration of electronic fences for multiple different risk dimensions. Based on the number of occurrences of each dangerous behavior type within areas with different regulatory levels, different adjustment values ​​are assigned to the intelligent agent group, thereby dynamically adjusting the electronic fence configuration based on the adjustment values.

[0054] Specifically, relevant data on all dangerous behaviors in areas with different regulatory levels are collected, including but not limited to video surveillance records, sensor data, personnel inspection reports, etc. These data are deeply integrated, and the image information in the video, the changes in environmental parameters detected by the sensor, and the abnormal situations recorded manually are uniformly classified and organized to determine the types of dangerous behaviors corresponding to areas with different regulatory levels. According to the different types of dangerous behaviors, the corresponding intelligent agent functions are defined respectively. The operation risk intelligent agent is responsible for monitoring and handling risks related to human operations, such as identifying illegal operating behaviors and adjusting the alarm sensitivity of the electronic fence; the environmental risk intelligent agent is used to monitor changes in environmental parameters. When abnormal temperature and humidity, leaks of harmful gases, etc. are detected, the ventilation, protection and other configurations of the electronic fence are automatically adjusted; the item risk intelligent agent monitors the status of the goods. When problems such as damage or displacement of the goods are found, the protection range and monitoring intensity of the electronic fence are adjusted.

[0055] Furthermore, a corresponding intelligent agent is created for each type of dangerous behavior, and the created intelligent agent is deployed in the warehouse safety management system, so that it can obtain relevant data of areas with different supervision levels in real time and establish a communication connection with the electronic fence equipment to adjust the electronic fence configuration.

[0056] Furthermore, statistics are collected on the number of occurrences of each type of dangerous behavior within different regulatory levels within a certain time period. Based on the number of occurrences of each dangerous behavior type and its potential harm to warehouse safety, different adjustment values ​​are assigned to the intelligent agent group. For example, dangerous behavior types with higher occurrences and greater potential harm are assigned higher adjustment values. Based on these assigned adjustment values, the intelligent agent group dynamically adjusts the configuration of the electronic fence.

[0057] In one embodiment of the present application, when a dangerous behavior alarm occurs in a supervision area, the associated supervision area corresponding to the supervision area is determined based on the spatial location of the electronic fence. The electronic fence of the associated supervision area is configured and adjusted according to the characteristics of the dangerous behavior.

[0058] Specifically, the spatial location information of all electronic fences within the regulatory area is centrally managed, including data such as the fence's coordinates, coverage, and boundary shape. Based on the spatial location of the electronic fences, the relationships between regulatory areas are analyzed. Adjacency is considered to determine which regulatory areas are directly connected in physical space. Traffic paths are analyzed to identify the routes of people and goods within the storage area, identifying regulatory areas that may pose risks due to the movement of people or goods.

[0059] When a dangerous behavior alarm occurs in a supervision area, the spatial location information of the alarm electronic fence is obtained. With the alarm location as the center, the supervision area associated with the alarm area is screened out from the electronic fence spatial location database. The characteristics of the alarmed dangerous behavior are analyzed, including the type, severity, scope of impact, development trend, etc. of the dangerous behavior. According to the obtained dangerous behavior characteristics, the corresponding adjustment strategy is determined in the electronic fence configuration adjustment database, and the formulated configuration adjustment strategy is sent to the electronic fence system of the associated supervision area. The electronic fence system automatically adjusts the relevant configuration parameters according to the received instructions. Among them, the electronic fence configuration adjustment database in the embodiment of the present application includes different dangerous characteristics, and also includes electronic fence adjustment parameters corresponding to different dangerous characteristics.

[0060] Figure 2 This is a schematic diagram of a dangerous behavior monitoring device based on an electronic fence provided in an embodiment of the present application. Figure 2 As shown, the dangerous behavior supervision device 200 based on the electronic fence includes: at least one processor 201; and a memory 202 in communication with the at least one processor 201; wherein the memory 202 stores instructions that can be executed by the at least one processor 201, and the instructions are executed by the at least one processor 201 so that the at least one processor 201 can: divide the storage area into multiple supervision areas based on the layout information and product information of the storage area, and divide the electronic fences based on the danger levels corresponding to the multiple supervision areas. Obtain the monitoring information corresponding to each electronic fence, and determine the severity level of the dangerous behavior corresponding to the multiple supervision areas based on the monitoring information. When the severity level of the dangerous behavior does not meet the preset level conditions, the supervision area is divided into supervision levels based on the location where the dangerous behavior occurs. Based on the supervision level, the electronic fences of the supervision area are dynamically adjusted to perform hierarchical supervision of the supervision area.

[0061] The embodiment of the present application provides a non-volatile computer storage medium storing computer executable instructions, wherein the computer executable instructions are configured to: divide the storage area into multiple supervision areas based on the layout information and product information of the storage area, and perform electronic fence division based on the danger levels corresponding to the multiple supervision areas. Obtain monitoring information corresponding to each electronic fence, and determine the severity level of dangerous behaviors corresponding to the multiple supervision areas based on the monitoring information. When the severity level of the dangerous behavior does not meet the preset level conditions, the supervision area is divided into supervision levels based on the location where the dangerous behavior occurs. Based on the supervision level, the electronic fences of the supervision area are dynamically adjusted to perform hierarchical supervision of the supervision area.

[0062] The various embodiments in this application are described in a progressive manner. Similar portions between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from the other embodiments. In particular, the device, apparatus, and non-volatile computer storage medium embodiments are generally similar to the method embodiments, so their descriptions are relatively simple. For relevant portions, refer to the descriptions of the method embodiments.

[0063] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. It will be apparent to those skilled in the art that various modifications and variations may be made to the embodiments of the present application. However, such modifications or substitutions do not deviate from the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

1. A dangerous behavior supervision method based on electronic fence, characterized in that: The method comprises: Based on the layout information and product information of the storage area, the storage area is divided into multiple supervision areas, and electronic fences are divided based on the danger levels corresponding to the multiple supervision areas; Acquiring monitoring information corresponding to each of the electronic fences, and determining the severity levels of dangerous behaviors corresponding to the plurality of supervision areas based on the monitoring information; If the severity level of the dangerous behavior does not meet the preset level conditions, the supervision area is divided into supervision levels based on the location where the dangerous behavior occurs; Based on the supervision level, the electronic fence of the supervision area is dynamically adjusted to perform hierarchical supervision on the supervision area.

2. The method for monitoring dangerous behaviors based on electronic fences according to claim 1, characterized in that: The method of dividing the storage area into multiple supervision areas based on the layout information and product information of the storage area, and performing electronic fence division based on the danger levels corresponding to the multiple supervision areas, specifically includes: Based on the layout information of the storage area, a first supervision area distribution is obtained; wherein the first supervision area distribution includes a functional area type, a functional area location, a functional area area, and a functional area shape; Based on the product information, the stored goods are classified and a second supervision area distribution corresponding to the same type of goods is determined in the product storage area; wherein the second supervision area distribution includes the type of stored goods, storage location, and storage area; Obtaining a first risk level corresponding to the first regulatory area distribution according to the first regulatory area distribution and the first distribution weight; Obtaining a second risk level corresponding to the second regulatory area distribution according to the second regulatory area distribution and the second distribution weight; Based on the first danger level and the second danger level, the electronic fence supervision level corresponding to each of the supervision areas is determined, and the electronic fence is divided.

3. The method for monitoring dangerous behaviors based on electronic fences according to claim 1, characterized in that: Determining the severity levels of dangerous behaviors corresponding to the plurality of supervision areas based on the monitoring information specifically includes: By using a preset deep learning model, the monitoring information is subjected to real-time risk feature extraction; wherein the risk feature includes the risk behavior type and the number of times the risk behavior occurs corresponding to each risk behavior type; Matching the extracted hazard features with time information and spatial information to construct a spatiotemporal hazard feature dataset; wherein the spatiotemporal hazard feature dataset includes at least a hazard occurrence timestamp and a hazard occurrence location; Based on the spatiotemporal hazard feature data set, determining a corresponding hazard level weight in a preset hazard level setting table; According to the spatiotemporal hazard feature data set and the hazard degree weight, the severity levels of dangerous behaviors corresponding to the plurality of supervision areas are determined.

4. The method for monitoring dangerous behaviors based on electronic fences according to claim 3, characterized in that: After extracting the danger features of the monitoring information in real time by using the preset deep learning model, the method further includes: In the case where the dangerous feature falls within the preset alarm feature range, determining a bounding box of the feature target in the video frame of the monitoring information; Obtain a target image through the bounding box of the feature target, and track the behavior of the feature target on the target image to obtain a motion trajectory of the feature target; According to the motion trajectory of the characteristic target, a predicted motion trend is obtained; According to the predicted movement trend, the dangerous consequence level corresponding to the dangerous feature is determined, so as to control the electronic fence to issue an alarm according to the dangerous consequence level.

5. The method for monitoring dangerous behaviors based on electronic fences according to claim 3 is characterized in that: Determining the severity levels of dangerous behaviors corresponding to the plurality of supervision areas respectively based on the spatiotemporal dangerous feature dataset and the dangerous degree weights specifically includes: Determining the spatiotemporal hazard characteristic data corresponding to each of the regulatory areas; Assigning type weights based on the dangerous behavior types corresponding to the spatiotemporal dangerous feature data; and, assigning a weight to each dangerous behavior based on the number of times the dangerous behavior occurs corresponding to each of the dangerous behavior types; The severity level of the dangerous behavior corresponding to the supervision area is obtained according to the spatiotemporal risk characteristic data, the type weight and the number weight.

6. The method for monitoring dangerous behaviors based on electronic fences according to claim 1, characterized in that: When the severity level of the dangerous behavior does not meet the preset level conditions, the supervision area is divided into supervision levels based on the location where the dangerous behavior occurs, specifically including: Marking the location where the dangerous behavior occurred on the electronic map corresponding to the storage area; Determining the scope of risk impact based on the risk behavior type corresponding to the risk behavior; Taking the marked point as the origin, draw the risk impact range corresponding to each of the risky behaviors in the supervision area; The hazard impact ranges are regionally superimposed to divide the supervision areas into supervision levels based on the superposition results.

7. The method for monitoring dangerous behaviors based on electronic fences according to claim 1, characterized in that: The dynamically regulating the electronic fence of the supervision area based on the supervision level to perform hierarchical supervision of the supervision area specifically includes: Determine the types of dangerous behaviors corresponding to different regulatory levels; wherein the types of dangerous behaviors include at least operational hazards and environmental hazards; and, constructing an agent group based on the dangerous behavior type; wherein the agent group is used to adjust the configuration of the electronic fence of multiple different dangerous dimensions; Based on the number of occurrences of each of the dangerous behavior types in the different supervision level areas, different adjustment values ​​are assigned to the intelligent agent groups, so as to dynamically regulate the electronic fence configuration based on the adjustment values.

8. The method for monitoring dangerous behaviors based on electronic fences according to claim 1, characterized in that: After performing hierarchical supervision on the supervision area, the method further includes: In the event of a dangerous behavior alarm in the supervision area, determining an associated supervision area corresponding to the supervision area based on the spatial position of the electronic fence; The electronic fence of the associated supervision area is configured and adjusted according to the characteristics of the dangerous behavior.

9. A dangerous behavior monitoring device based on an electronic fence, characterized in that: The device comprises a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the device is triggered to execute the method according to any one of claims 1 to 8.

10. A non-volatile computer storage medium storing computer executable instructions, characterized in that: The computer executable instructions can execute the method according to any one of claims 1 to 8.

Images