Electronic fence-based personnel platform management method and system

Abstract

This invention discloses a personnel platform management method and system based on electronic fences, relating to the field of management and scheduling technology. The method includes acquiring original images within the electronic fence area, preprocessing the original images to obtain enhanced images, and inputting the enhanced images into a target model to obtain target data. The target data includes accident type, accident location, accident quantity, and accident level. The method also involves acquiring the work status data corresponding to each security personnel, analyzing the work status data and the target data, filtering out security personnel who meet the criteria to obtain security instructions, and sending the security instructions to the security personnel's terminals. By improving image quality through image preprocessing and accurately identifying accident information in conjunction with the target model, the method optimizes task allocation based on the security personnel's status, achieving rapid response and efficient decision-making, and significantly improving emergency response efficiency.

Description

Technical Field

[0001] This invention belongs to the field of management and scheduling technology, specifically relating to a personnel platform management method and system based on electronic fences. Background Technology

[0002] An electronic fence is a virtual boundary management method based on positioning and communication technologies, commonly used in security systems for the regional control of personnel, vehicles, or equipment. Its principle is to delineate a specific area on an electronic map or in a real-world scenario. When a monitored object enters, leaves, or stays in the area for more than a preset time, the system automatically triggers an alarm or notification. Through technologies such as GPS, Bluetooth, Wi-Fi, and UWB, electronic fences can achieve precise monitoring both indoors and outdoors. In security applications, it not only prevents unauthorized personnel from entering sensitive areas but can also be used to monitor whether security personnel are patrolling according to regulations, preventing them from leaving their posts or gathering abnormally, thereby improving the intelligence and efficiency of security management.

[0003] In low-light or foggy environments, existing security systems lack accuracy and timeliness in image analysis, making it difficult to accurately and quickly identify the time, type, location, and severity of incidents within electronic fence areas. Consequently, the security personnel management system lacks reliable data support and fails to comprehensively consider the real-time status and qualifications of security personnel, relying solely on physical distance for allocation. This results in uneven task distribution, low response efficiency, and delayed personnel management and scheduling. There are significant shortcomings in emergency response capabilities, necessitating optimization and upgrades to enhance emergency handling capabilities. Summary of the Invention

[0004] The purpose of this invention is to solve the problem of slow response and low efficiency in personnel management and scheduling caused by the lack of reliable data in personnel management systems, and to propose a personnel platform management method based on electronic fences.

[0005] In a first aspect of this invention, a personnel platform management method based on electronic fences is first proposed, the method comprising: Acquire the original image within the electronic fence area, and preprocess the original image to obtain an enhanced image; The enhanced image is input into the target model to obtain target data; the target data includes accident type, accident location, accident number, and accident level; The system acquires the work status data of each security personnel, analyzes the work status data with the target data, filters out security personnel who meet the conditions, obtains security instructions, and sends the security instructions to the security personnel's terminals. The work status data includes whether the personnel are on duty, their real-time location, age, and personnel level. The security instructions are assigned to the qualified security personnel through the terminals.

[0006] Optionally, the original image is preprocessed to obtain an enhanced image, including: Extract the green channel and blue channel pixel data from the original image; set a fixed average gray value according to the Gray World algorithm; The global average intensity value of the green channel is calculated by iterating through all pixels of the green channel, and the global average intensity value of the blue channel is calculated by iterating through all pixels of the blue channel. The gain ratio of the green channel is calculated based on the average gray value and the average intensity value of the green channel; the gain ratio of the blue channel is calculated based on the average gray value and the average intensity value of the blue channel. Iterate through each pixel in the green channel and multiply each green channel pixel by the green channel gain ratio to obtain the corrected green channel pixel value; iterate through each pixel in the blue channel and multiply each blue channel pixel by the blue channel gain ratio to obtain the corrected blue channel pixel value. The enhanced image is calculated based on the corrected green channel pixel values ​​and the corrected blue channel pixel values.

[0007] Optionally, an enhanced image is calculated based on the corrected green channel pixel values ​​and the corrected blue channel pixel values, specifically including: The global average intensity value is calculated by traversing the corrected green channel pixels; the global average intensity value is calculated by traversing the red channel pixels in the original image. The compensated red channel pixel value is calculated for each pixel in the red channel of the original image; The average value of the red channel pixels after traversal compensation is calculated, the average value of the green channel pixels after traversal correction is calculated, and the average value of the blue channel pixels after traversal correction is calculated. , and The average grayscale reference of the three channels was calculated. ; Multiply each pixel of the channel by Obtain the corrected R pixel value; multiply each pixel of the channel by Obtain the corrected G pixel value; multiply each pixel of the channel by Obtain the corrected B pixel value; The corrected R, G, and B pixel values ​​are combined according to their corresponding pixel coordinates to obtain the enhanced image. Optionally, the target model is an improvement based on YOLOv11, and the improvement includes: In the YOLOv11 backbone network, the YOLOv11 Layer 2 C3K2 module, Layer 4 C3K2 module, and Layer 6 C3K2 module are replaced with ZAWRCNet modules; In the YOLOv11 backbone network, the YOLOv11 Layer 7 Conv and Layer 8 C3K2 modules are removed; The ZAWRCNet module works as follows: Obtain the input features, input the input features into the Conv module to obtain the first convolutional feature, and split the first convolutional feature to obtain the first channel feature and the second channel feature; The first channel feature is input into the Star Res-CAA Block module to obtain the first initial feature; the first initial feature is input into the Star Res-CAA Block module to obtain the second initial feature; The Star Res-CAA Block module comprises a dimensionality reduction layer, a feature extraction layer, and a dimensionality increase layer in sequence. The first convolutional feature is input into the Conv module to obtain the second convolutional feature; The first convolutional feature is sequentially input into the Conv module, DWConv module, and Point-wise Conv module to obtain the third convolutional feature; The first target feature is obtained by concatenating the first channel feature, the second channel feature, the first initial feature, the second convolutional feature, the second initial feature, and the third convolutional feature. The first target feature is then input into the Conv module to obtain the second target feature. The second target feature is used as the final output image of the ZAWRCNet module.

[0008] Optionally, the work status data and the target data are analyzed to filter out security personnel who meet the conditions and receive security instructions, including: The distance between the real-time location of each security personnel and the accident location is calculated based on the target data and the working status data. Based on the work status data and the distance data, the priority value of the security personnel relative to other personnel is calculated, and the task allocation order and task content of the security personnel in the security instructions are adjusted according to the priority value. The system tracks the dynamic location of security personnel in real time, updates their distance data, and dynamically adjusts task allocation.

[0009] In a second aspect of this invention, a personnel platform management system based on electronic fences is proposed, comprising an image processing module, a target data acquisition module, and a task allocation module, wherein: The image processing module is used to acquire the original image within the electronic fence area and preprocess the original image to obtain an enhanced image; The task allocation module is used to acquire the work status data of each security personnel, analyze the work status data with the target data, filter out security personnel who meet the conditions, obtain security instructions, and send the security instructions to the security personnel's terminals; the work status data includes whether the personnel are on duty, real-time location, age, and personnel level; the security instructions are used to allocate tasks to security personnel who meet the conditions through the terminals.

[0010] Optionally, the image processing module includes a grayscale value setting module, an average intensity first calculation module, a gain ratio calculation module, a pixel correction first module, and an enhanced image acquisition module, wherein: The grayscale setting module is used to extract green channel pixel data and blue channel pixel data from the original image; and set a fixed average grayscale value according to the Gray World algorithm. The first average intensity calculation module is used to calculate the global average intensity value of the green channel by traversing all pixels of the green channel, and to calculate the global average intensity value of the blue channel by traversing all pixels of the blue channel. The gain ratio calculation module is used to calculate the gain ratio of the green channel based on the average gray value and the average intensity value of the green channel; and to calculate the gain ratio of the blue channel based on the average gray value and the average intensity value of the blue channel. The pixel correction first module is used to iterate through each pixel in the green channel, multiply each pixel in the green channel by the gain ratio of the green channel to obtain the corrected green channel pixel value; and to iterate through each pixel in the blue channel, multiply each pixel in the blue channel by the gain ratio of the blue channel to obtain the corrected blue channel pixel value. The enhanced image acquisition module is used to calculate the enhanced image based on the corrected green channel pixel values ​​and the corrected blue channel pixel values.

[0011] Optionally, the enhanced image acquisition module includes a second average intensity calculation module, a compensation pixel calculation module, an average grayscale reference acquisition module, a second pixel correction module, and a combination module, wherein: The second average intensity calculation module is used to calculate the global average intensity value by traversing the corrected green channel pixels and calculating the global average intensity value by traversing the red channel pixels in the original image. The compensation pixel calculation module is used to calculate the compensated red channel pixel value for each pixel in the red channel of the original image. The average grayscale reference acquisition module is used to calculate the average value of the red channel pixels after compensation, the average value of the green channel pixels after correction, and the average value of the blue channel pixels after correction, based on... , and The average grayscale reference of the three channels was calculated. ; The second pixel correction module is used to multiply each pixel of the channel by Obtain the corrected R pixel value; multiply each pixel of the channel by Obtain the corrected G pixel value; multiply each pixel of the channel by Obtain the corrected B pixel value; The combination module is used to combine the corrected R, G, and B pixel values ​​according to their corresponding pixel coordinates to obtain an enhanced image.

[0012] Optionally, the target data acquisition module includes an improvement based on YOLOv11, wherein: In the YOLOv11 backbone network, the YOLOv11 Layer 2 C3K2 module, Layer 4 C3K2 module, and Layer 6 C3K2 module are replaced with ZAWRCNet modules; In the YOLOv11 backbone network, the YOLOv11 Layer 7 Conv and Layer 8 C3K2 modules are removed; The ZAWRCNet module works as follows: Obtain the input features, input the input features into the Conv module to obtain the first convolutional feature, and split the first convolutional feature to obtain the first channel feature and the second channel feature; The first channel feature is input into the Star Res-CAA Block module to obtain the first initial feature; the first initial feature is input into the Star Res-CAA Block module to obtain the second initial feature; The Star Res-CAA Block module comprises a dimensionality reduction layer, a feature extraction layer, and a dimensionality increase layer in sequence. The Star Res-CAA Block module comprises a dimensionality reduction layer, a feature extraction layer, and a dimensionality increase layer in sequence. The first convolutional feature is input into the Conv module to obtain the second convolutional feature; The first convolutional feature is sequentially input into the Conv module, DWConv module, and Point-wise Conv module to obtain the third convolutional feature; The first target feature is obtained by concatenating the first channel feature, the second channel feature, the first initial feature, the second convolutional feature, the second initial feature, and the third convolutional feature. The first target feature is then input into the Conv module to obtain the second target feature. The second target feature is used as the final output image of the ZAWRCNet module.

[0013] Optionally, the task allocation module includes a distance calculation module, a priority module, and a dynamic adjustment module, wherein: The distance calculation module is used to calculate the distance data between the real-time location of each security personnel and the accident location based on the target data and the working status data. The priority module is used to calculate the priority value of security personnel relative to other personnel based on the work status data and the distance data, and adjust the task allocation order and task content of security personnel in the security instructions according to the priority value; The dynamic adjustment module is used to track the dynamic location of security personnel in real time, update their distance data, and dynamically adjust task allocation.

[0014] The beneficial effects of this invention are: This invention proposes a personnel platform management method based on electronic fences. It improves image quality through image preprocessing, providing clear data for subsequent analysis; it quickly and accurately identifies key accident information based on the target model, facilitating reliable data support for subsequent task allocation; and it optimizes task allocation based on the status and distance of security personnel, achieving efficient and intelligent scheduling and dynamic management, effectively shortening task allocation time and accident handling time, improving emergency response efficiency, and enhancing emergency handling capabilities. Attached Figure Description

[0015] The invention will now be further described with reference to the accompanying drawings.

[0016] Figure 1 A flowchart illustrating a personnel platform management method based on an electronic fence, provided as an embodiment of the present invention; Figure 2 A structural diagram of a personnel platform management method based on electronic fence provided in an embodiment of the present invention; Figure 3 This is a structural diagram of the ZAWRCNet module provided in an embodiment of the present invention; Figure 4 This is a framework diagram of the Star Res-CAA Block module provided in an embodiment of the present invention; Figure 5 This is a block diagram of the Res-CAA module provided in an embodiment of the present invention; Figure 6 A flowchart of the Star Block module provided in an embodiment of the present invention; Figure 7 This is a framework diagram of a personnel platform management system based on electronic fences, provided for an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and B can represent: A alone, A and B simultaneously, and B alone. Furthermore, descriptions involving "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" can explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0018] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] This invention provides a personnel platform management method based on electronic fences. See also... Figure 1 , Figure 1 A flowchart illustrating a personnel platform management method based on an electronic fence, provided as an embodiment of the present invention. The method includes the following steps: S101, acquire the original image within the electronic fence area, and preprocess the original image to obtain the enhanced image; S102, The enhanced image is input into the target model to obtain the target data; S103: Obtain the work status data corresponding to each security personnel, analyze the work status data and target data, filter out the security personnel who meet the conditions, obtain security instructions, and send the security instructions to the security personnel's terminals. The target data includes accident type, accident location, and accident level; the work status data includes whether the personnel are on duty, their real-time location, age, and personnel level.

[0020] The personnel platform management method based on electronic fence provided by this invention improves image quality through image preprocessing, providing clear data for subsequent analysis; it quickly and accurately identifies key accident information based on the target model, facilitating effective data for subsequent task allocation; and it optimizes task allocation based on the status and distance of security personnel, achieving efficient and intelligent scheduling and dynamic management, effectively shortening task allocation time and accident handling time, improving emergency response efficiency, and enhancing emergency handling capabilities.

[0021] In one implementation, security personnel within the electronic fence area are equipped with terminals, such as mobile phones. On weekdays, security personnel clock in and out using the terminals, and during work hours, pre-assigned tasks are distributed to the corresponding security personnel via the terminals. Based on distance data and work status data, security instructions are adjusted to assign tasks to personnel in descending order.

[0022] In one implementation, surveillance cameras installed in the target area acquire real-time image data of the surrounding environment to meet the need for timely response to emergencies; the distance data between the target data and the personnel in the security instructions is calculated to determine the fastest path distance for the security personnel to move from their current location to the location of the accident in the target data.

[0023] In one embodiment, preprocessing the original image to obtain an enhanced image includes: Extract the green channel and blue channel pixel data from the original image; set a fixed average gray value according to the Gray World algorithm; The global average pixel value of the green channel is calculated by iterating through all pixels in the green channel. The global average pixel value of the blue channel is calculated by iterating through all pixels in the blue channel. ; Based on average grayscale value and average pixel value of green channel The gain ratio of the green channel was calculated; based on the average grayscale value and the average pixel value of the blue channel. The gain ratio of the blue channel was calculated. Iterate through each pixel in the green channel and multiply each green channel pixel by the green channel gain ratio to obtain the corrected green channel pixel value. Iterate through each pixel in the blue channel, multiplying each blue channel pixel by the blue channel gain ratio to obtain the corrected blue channel pixel value. ; Based on the corrected green channel pixel values and the corrected blue channel pixel values The enhanced image is calculated.

[0024] In one implementation, the global average pixel value is obtained by iterating through all pixels in the corresponding channel, summing the pixels, and dividing by the total number of pixels. In one implementation, low-light environments can easily lead to image intensity imbalance. The blue (B) and green (G) channels in the original image carry the most information, so the two channels need to be equalized first. By extracting and adjusting the gain of the pixel data of the green and blue channels in the original image, the global brightness of the image is made closer to the set fixed average gray value, thereby effectively correcting the problem of uneven brightness. By combining the corrected green and blue channels with the original red channel, a uniform and clear enhanced image is generated, improving image quality.

[0025] In one embodiment, based on the corrected green channel pixel values and the corrected blue channel pixel values The enhanced image is calculated, including: After traversing and correcting the green channel pixels Calculate its global average pixel value Calculate the global average pixel value by traversing the red channel pixels in the original image. ; The compensated red channel pixel value is calculated for each pixel in the original image's red channel. Calculate the average value of the red channel pixels after traversing the compensated values. After traversing and correcting the green channel pixels Calculate the average value Traversing the corrected blue channel pixel values Calculate the average value ,according to , and The average grayscale reference of the three channels was calculated. ; right Each pixel of the channel is multiplied / Obtain the corrected R pixel value; for Each pixel of the channel is multiplied / Obtain the corrected G pixel value; for Each pixel of the channel is multiplied / Obtain the corrected B pixel value; combine the corrected R, G, and B pixel values ​​according to their corresponding pixel coordinates to obtain the enhanced image.

[0026] In one implementation, the compensated red channel pixel value is calculated for each pixel in the original image: in, To compensate for the pixel value of the red channel at the nth pixel position, This represents the pixel value of the original red channel at the nth pixel position. To correct the pixel value of the green channel at the nth pixel position; The compensation coefficient is determined to be 1 based on historical experimental data, ensuring that the compensation range is reasonable and that no oversaturation occurs.

[0027] In one implementation, the attenuation rate of light in low-light environments is positively correlated with wavelength; the longer the wavelength, the faster the attenuation. The attenuation order of the three visible light sources is red light > green light > blue light. The attenuation rate of green light is between that of red and blue light, and it is the wavelength that the human eye is most sensitive to. The green channel can retain sufficient penetration depth while reducing information redundancy caused by scattering, thus carrying the most effective structural details of the image. Therefore, the green channel has the highest information integrity and the least distortion in the image, making it suitable as the benchmark for calculating red channel compensation. However, the blue channel has a lot of scattering redundancy, and if used as the benchmark, it will introduce blue pollution, which will lead to color distortion after red channel compensation.

[0028] In one implementation, the average value of the compensated red channel pixel values ​​is calculated. Corrected green channel pixels Calculate the average value and corrected blue channel pixel values Calculate the average value The process is as follows:

[0029] Where x represents the image width and y represents the image height; This represents the pixel value in the m-th row and n-th column of the red channel; This represents the pixel value in the m-th row and n-th column of the green channel; This represents the pixel value in the m-th row and n-th column of the blue channel; It is a three-channel average grayscale reference; It is an enhanced image obtained through processing.

[0030] In one implementation, the average values ​​of the corrected green and blue channels and the compensated red channel are calculated separately, and a unified average grayscale benchmark is determined by combining the three channel data, so that the brightness of each pixel is gradually adjusted to a more balanced state; the final enhanced image significantly improves the brightness coordination between color channels, enhances the uniformity and clarity of the overall image, and provides better data support for subsequent input to the target model.

[0031] In one embodiment, the target model is an improvement based on YOLOv11, and the improvements include: In the YOLOv11 backbone network, the YOLOv11 Layer 2 C3K2 module, Layer 4 C3K2 module, and Layer 6 C3K2 module are replaced with ZAWRCNet modules; In the YOLOv11 backbone network, the YOLOv11 Layer 7 Conv and Layer 8 C3K2 modules are removed; The ZAWRCNet module works as follows: Obtain the input features, input the input features into the Conv module to obtain the first convolutional feature, and split the first convolutional feature to obtain the first channel feature and the second channel feature; The Star Res-CAA Block module consists of a dimensionality reduction layer, a feature extraction layer, and a dimensionality increase layer. The first channel feature is input into the Star Res-CAA Block module to obtain the first initial feature; the first initial feature is input into the Star Res-CAA Block module to obtain the second initial feature; The first convolutional feature is input into the Conv module to obtain the second convolutional feature; The first convolutional feature is sequentially input into the Conv module, DWConv module, and Point-wise Conv module to obtain the third convolutional feature; The first target feature is obtained by concatenating the first channel feature, the second channel feature, the first initial feature, the second convolutional feature, the second initial feature, and the third convolutional feature. The first target feature is then input into the Conv module to obtain the second target feature. The second target feature is used as the final output image of the ZAWRCNet module.

[0032] In one implementation, see [link to implementation details]. Figure 2 , Figure 2 A model structure diagram of a personnel platform management method based on electronic fence provided in an embodiment of the present invention; The target model optimizes the detection layer structure by removing the original detection layer for large targets and adding a high-resolution detection layer to meet the target detection requirements. It introduces the ZAWRCNet feature extraction module, which combines a multi-branch structure with star-shaped operations and utilizes the residual context anchor attention mechanism to achieve deep fusion of high- and low-dimensional features and contextual information. It integrates bounding box quality information into the classification task to improve the synergy between classification and regression, effectively improving the accuracy of target localization.

[0033] In one implementation, see [link to implementation details]. Figure 3 , Figure 3 This is a structural diagram of the ZAWRCNet module provided in an embodiment of the present invention; Figure 3 This is the structure of the ZAWRCNet module. Figure 3The Star Res-CAA Block integrates the Star Block module and the Res-CAA mechanism to efficiently enhance the feature representation of small targets. The Star Block maps features to a high-dimensional space through star-shaped elements, fusing semantic and detailed information. The Res-CAA mechanism uses strip convolution to capture spatial correlations, combines channel attention to suppress background noise, and uses residual connections to pass on the original features. The combination of the two enhances the integrity of small target features, while the lightweight design controls computational costs and improves the model's adaptability to blurred small targets and complex scenes.

[0034] Alternatively, the Star Res-CAA Block module can be replaced by the Bottleneck module.

[0035] In one implementation, Figure 4 This is a framework diagram of the Star Res-CAA Block module provided in an embodiment of the present invention; the working principle of the Star Res-CAA Block module is as follows: The first channel feature is input into the DWConv module to obtain the first original feature. The first original feature is then input into the fully connected layer FC to obtain the second and third original features. The second and third original features are fused to obtain the fourth original feature. The fourth original feature is input into the Res-CAA module to obtain the fifth original feature. The fifth original feature is input into the fully connected layer FC to obtain the sixth original feature. The sixth original feature is then input into the DWConv module to obtain the seventh original feature. The first channel feature and the seventh original feature are then added together to output the first initial feature.

[0036] In one implementation, see [link to implementation details]. Figure 5 , Figure 5 The diagram below shows the block diagram of the Res-CAA module provided in this embodiment of the invention. The working principle of the Res-CAA module is as follows: the fourth original feature is subjected to average pooling to obtain the first enhanced feature. The first enhanced feature is then sequentially input into the Conv module, DWConv module, DWConv module and Conv module to obtain the second enhanced feature. The second enhanced feature is then processed by the Sigmoid function to obtain the third enhanced feature. The fourth original feature and the third enhanced feature are added together to output the fifth original feature.

[0037] In one implementation, see [link to implementation details]. Figure 6 , Figure 6 The flowchart of the Star Block module provided in this embodiment of the invention is shown. The Star Block module efficiently integrates multi-dimensional features through multi-branch and star-shaped operations, enhances feature expression and extraction capabilities, effectively handles small target detection, improves detection accuracy, and is designed to be lightweight and adaptable to real-time task requirements.

[0038] In one implementation, experiments were conducted using the UCF-Crime dataset. The software systems used in the experiments included Windows 11, Python 3.8, PyTorch 2.0.0, and CUDA 11.8; the hardware configuration consisted of an Intel Xeon Gold 5117 processor and an NVIDIA Tesla V100-PCIE graphics card. Table 1. Detection results for different accident types based on the UCF-Crime dataset.

[0039] As can be seen from Table 1, the target model has improved the detection accuracy, recall rate and mAP@0.75 of the accident type identification compared with the previous YOLOv11. Therefore, the target model is more accurate and efficient in identifying accident types in the target area under low light conditions.

[0040] In one embodiment, analyzing the work status data and the target data to filter out security personnel who meet the criteria and receive security instructions includes: The distance between the real-time location of each security personnel and the accident location is calculated based on the target data and work status data. Based on work status data and distance data, the priority value of security personnel relative to other personnel is calculated, and the task allocation order and task content of security personnel in security instructions are adjusted according to the priority value. The system tracks the dynamic location of security personnel in real time, updates their distance data, and dynamically adjusts task allocation.

[0041] In one implementation, the incident set Every accident There is a specific location Accident type and accident level (0 Mild - 1 Severe); Personnel gathering Each person Real-time location Are they on duty? Personnel Level Higher level is better, and the current task load of personnel is also considered. (0 idle, 1 fully loaded), estimated average speed is ; Distance data is obtained through calculation:

[0042] in, It is the distance between the accident location and the real-time location of the personnel. This is the estimated time required for security personnel to reach the location of the incident; As a priority value, , , and The weights are set based on historical experimental data and can be adjusted manually.

[0043] In one implementation, task allocation is dynamically adjusted by updating personnel locations and ETAs every T time interval. If the ETA of currently assigned security personnel for an incident increases beyond a threshold, tasks are reassigned; if the incident level... If the increase is significant, a rescheduling will be triggered immediately.

[0044] Based on the same inventive concept, this invention also provides a personnel platform management system based on electronic fences. See also Figure 7 , Figure 7 A schematic diagram of a personnel platform management system based on an electronic fence, provided for an embodiment of the present invention, includes an image processing module, a target data acquisition module, and a task allocation module, wherein: The image processing module is used to acquire the original image within the electronic fence area and preprocess the original image to obtain the enhanced image; the target area is the area defined by the electronic fence. The target data acquisition module is used to input the enhanced image into the target model to obtain target data; the target data includes accident type, accident location, accident number, and accident level; The task allocation module is used to acquire the work status data of each security personnel, analyze the work status data with the target data, filter out the security personnel who meet the conditions, obtain security instructions, and send the security instructions to the security personnel's terminals. The work status data includes whether the personnel are on duty, their real-time location, age, and personnel level. The security instructions are assigned to the security personnel who meet the conditions through the terminals.

[0045] The personnel platform management system based on electronic fences provided by this invention improves image quality through image preprocessing, providing clear data for subsequent analysis; it quickly and accurately identifies key accident information based on the target model, facilitating effective data for subsequent task allocation; and it optimizes task allocation based on the status and distance of security personnel, achieving efficient and intelligent scheduling and dynamic management, effectively shortening task allocation time and accident handling time, improving emergency response efficiency, and enhancing emergency handling capabilities.

[0046] In one embodiment, the image processing module includes a grayscale value setting module, an average pixel first calculation module, a gain ratio calculation module, a pixel correction first module, and an enhanced image acquisition module, wherein: The grayscale setting module is used to extract green channel and blue channel pixel data from the original image; and set a fixed average grayscale value according to the Gray World algorithm. The first average pixel calculation module is used to calculate the global average pixel value of the green channel by iterating through all pixels in the green channel. The global average pixel value of the blue channel is calculated by iterating through all pixels in the blue channel. ; The gain ratio calculation module is used to calculate the gain ratio based on the average grayscale value and the average pixel value of the green channel. The gain ratio of the green channel was calculated; based on the average grayscale value and the average pixel value of the blue channel. The gain ratio of the blue channel was calculated. The first pixel correction module iterates through each pixel in the green channel, multiplying each green channel pixel by the green channel's gain ratio to obtain the corrected green channel pixel value. Iterate through each pixel in the blue channel, multiplying each blue channel pixel by the blue channel gain ratio to obtain the corrected blue channel pixel value. ; The enhanced image acquisition module is used to obtain the corrected green channel pixel values. and the corrected blue channel pixel values The enhanced image is calculated.

[0047] In one embodiment, the enhanced image acquisition module includes a second average pixel calculation module, a compensated pixel calculation module, an average grayscale reference acquisition module, a second pixel correction module, and a combination module, wherein: The second average pixel calculation module is used to iterate through the corrected green channel pixels. Calculate its global average pixel value Calculate the global average pixel value by traversing the red channel pixels in the original image. The compensation pixel calculation module is used to calculate the compensated red channel pixel value for each pixel in the original image's red channel. ; The average grayscale reference acquisition module is used to calculate the average value by traversing the compensated red channel pixel values. After traversing and correcting the green channel pixels Calculate the average value Traversing the corrected blue channel pixel values Calculate the average value ,according to , and The average grayscale reference of the three channels was calculated. ; The second pixel correction module is used for... Each pixel of the channel is multiplied / Obtain the corrected R pixel value; for Each pixel of the channel is multiplied / Obtain the corrected G pixel value; for Each pixel of the channel is multiplied / Obtain the corrected B pixel value; The combination module is used to combine the corrected R, G, and B pixel values ​​according to their corresponding pixel coordinates to obtain an enhanced image.

[0048] In one embodiment, the target data acquisition module includes improvements based on YOLOv11, wherein: In the YOLOv11 backbone network, the YOLOv11 Layer 2 C3K2 module, Layer 4 C3K2 module, and Layer 6 C3K2 module are replaced with ZAWRCNet modules; In the YOLOv11 backbone network, the YOLOv11 Layer 7 Conv and Layer 8 C3K2 modules are removed; The ZAWRCNet module works as follows: Obtain the input features, input the input features into the Conv module to obtain the first convolutional feature, and split the first convolutional feature to obtain the first channel feature and the second channel feature; The first channel feature is input into the Star Res-CAA Block module to obtain the first initial feature; the first initial feature is input into the Star Res-CAA Block module to obtain the second initial feature; The Star Res-CAA Block module consists of a dimensionality reduction layer, a feature extraction layer, and a dimensionality increase layer. The first convolutional feature is input into the Conv module to obtain the second convolutional feature; The first convolutional feature is sequentially input into the Conv module, DWConv module, and Point-wise Conv module to obtain the third convolutional feature; The first target feature is obtained by concatenating the first channel feature, the second channel feature, the first initial feature, the second convolutional feature, the second initial feature, and the third convolutional feature. The first target feature is then input into the Conv module to obtain the second target feature. The second target feature is used as the final output image of the ZAWRCNet module.

[0049] In one embodiment, the task allocation module includes a distance calculation module, a priority module, and a dynamic adjustment module, wherein: The distance calculation module is used to calculate the distance between the real-time location of each security personnel and the accident location based on the target data and work status data. The priority module is used to calculate the priority value of security personnel relative to other personnel based on work status data and distance data, and adjust the task allocation order and task content of security personnel in security instructions according to the priority value; The dynamic adjustment module is used to track the dynamic location of security personnel in real time, update their distance data, and dynamically adjust task allocation.

[0050] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A personnel platform management method and system based on electronic fences, characterized in that, The method includes: Acquire the original image within the electronic fence area, and preprocess the original image to obtain an enhanced image; The enhanced image is input into the target model to obtain target data; the target data includes accident type, accident location, accident number, and accident level; The system acquires the work status data of each security personnel, analyzes the work status data with the target data, filters out security personnel who meet the conditions, obtains security instructions, and sends the security instructions to the security personnel's terminals. The work status data includes whether the personnel are on duty, their real-time location, age, and personnel level. The security instructions are assigned to the qualified security personnel through the terminals.

2. The personnel platform management method and system based on electronic fences according to claim 1, characterized in that, The enhanced image is obtained by preprocessing the original image, including: Extract the green channel and blue channel pixel data from the original image; set a fixed average gray value according to the Gray World algorithm; The global average pixel value of the green channel is calculated by iterating through all pixels in the green channel. The global average pixel value of the blue channel is calculated by iterating through all pixels in the blue channel. ; Based on the average grayscale value and the average pixel value of the green channel The gain ratio of the green channel is calculated; based on the average grayscale value and the average pixel value of the blue channel... The gain ratio of the blue channel was calculated. Iterate through each pixel in the green channel and multiply each green channel pixel by the green channel gain ratio to obtain the corrected green channel pixel value. Iterate through each pixel in the blue channel, multiplying each blue channel pixel by the blue channel gain ratio to obtain the corrected blue channel pixel value. According to the corrected green channel pixel values and the corrected blue channel pixel values The enhanced image is calculated.

3. The personnel platform management method and system based on electronic fences according to claim 2, characterized in that, Based on the corrected green channel pixel values and the corrected blue channel pixel values The enhanced image is calculated by traversing the corrected green channel pixels. Calculate its global average pixel value Calculate the global average pixel value by traversing the red channel pixels in the original image. ; The compensated red channel pixel value is calculated for each pixel in the original image's red channel. ; Calculate the average value of the red channel pixels after traversing the compensation. After traversing and correcting the green channel pixels Calculate the average value Traversing the corrected blue channel pixel values Calculate the average value ,according to , and The average grayscale reference of the three channels was calculated. ; Regarding the Each pixel of the channel is multiplied / Obtain the corrected R pixel value; for the Each pixel of the channel is multiplied / Obtain the corrected G pixel value; for the Each pixel of the channel is multiplied / Obtain the corrected B pixel value; The corrected R, G, and B pixel values ​​are combined according to their corresponding pixel coordinates to obtain the enhanced image.

4. The personnel platform management method and system based on electronic fences according to claim 1, characterized in that, The target model is an improvement based on YOLOv11, and the improvements include: In the YOLOv11 backbone network, the YOLOv11 Layer 2 C3K2 module, Layer 4 C3K2 module, and Layer 6 C3K2 module are replaced with ZAWRCNet modules; In the YOLOv11 backbone network, the YOLOv11 layer 7 Conv and layer 8 C3K2 modules are removed; The ZAWRCNet module works as follows: Obtain the input features, input the input features into the Conv module to obtain the first convolutional feature, and split the first convolutional feature to obtain the first channel feature and the second channel feature; The first channel feature is input into the Star Res-CAA Block module to obtain the first initial feature; the first initial feature is input into the Star Res-CAA Block module to obtain the second initial feature; The Star Res-CAA Block module comprises a dimensionality reduction layer, a feature extraction layer, and a dimensionality increase layer in sequence. The first convolutional feature is input into the Conv module to obtain the second convolutional feature; the first convolutional feature is then input into the Conv module, DWConv module, and Point-wise Conv module in sequence to obtain the third convolutional feature; The first target feature is obtained by concatenating the first channel feature, the second channel feature, the first initial feature, the second convolutional feature, the second initial feature, and the third convolutional feature. The first target feature is then input into the Conv module to obtain the second target feature. The second target feature is used as the final output image of the ZAWRCNet module.

5. The personnel platform management method and system based on electronic fences according to claim 1, characterized in that, The work status data is analyzed with the target data to filter out security personnel who meet the criteria and receive security instructions, including: The distance between the real-time location of each security personnel and the accident location is calculated based on the target data and the working status data. Based on the work status data and the distance data, the priority value of the security personnel relative to other personnel is calculated, and the task allocation order and task content of the security personnel in the security instructions are adjusted according to the priority value. The system tracks the dynamic location of security personnel in real time, updates their distance data, and dynamically adjusts task allocation.

6. A personnel platform management method and system based on electronic fences, characterized in that, The system includes an image processing module, a target data acquisition module, and a task allocation module, wherein: The image processing module is used to acquire the original image within the electronic fence area and preprocess the original image to obtain an enhanced image; The target data acquisition module is used to input the enhanced image into the target model to obtain target data; the target data includes accident type, accident location, accident number, and accident level; The task allocation module is used to acquire the work status data of each security personnel, analyze the work status data with the target data, filter out security personnel who meet the conditions, obtain security instructions, and send the security instructions to the security personnel's terminals; the work status data includes whether the personnel are on duty, real-time location, age, and personnel level; the security instructions are used to allocate tasks to security personnel who meet the conditions through the terminals.

7. The personnel platform management method and system based on electronic fences according to claim 6, characterized in that, The image processing module includes a grayscale value setting module, an average pixel first calculation module, a gain ratio calculation module, a pixel correction first module, and an enhanced image acquisition module, wherein: The grayscale setting module is used to extract green channel pixel data and blue channel pixel data from the original image; and set a fixed average grayscale value according to Gray World calculation. The first average pixel calculation module is used to calculate the global average pixel value of the green channel by traversing all pixels of the green channel. The global average pixel value of the blue channel is calculated by iterating through all pixels in the blue channel. ; The gain ratio calculation module is used to calculate the gain ratio based on the average grayscale value and the average pixel value of the green channel. The gain ratio of the green channel is calculated; based on the average grayscale value and the average pixel value of the blue channel... The gain ratio of the blue channel was calculated. The first pixel correction module is used to iterate through each pixel in the green channel and multiply each pixel by the gain ratio of the green channel to obtain the corrected green channel pixel value. Iterate through each pixel in the blue channel, multiplying each blue channel pixel by the blue channel gain ratio to obtain the corrected blue channel pixel value. ; The enhanced image acquisition module is used to obtain the corrected green channel pixel values. and the corrected blue channel pixel values The enhanced image is calculated.

8. The personnel platform management method and system based on electronic fence according to claim 7, characterized in that, The enhanced image acquisition module includes a second average pixel calculation module, a compensated pixel calculation module, an average grayscale reference acquisition module, a second pixel correction module, and a combination module, wherein: The second average pixel calculation module is used to iterate through the corrected green channel pixels. Calculate its global average pixel value Calculate the global average pixel value by traversing the red channel pixels in the original image. ; The compensation pixel calculation module is used to calculate the compensated red channel pixel value for each pixel in the original image's red channel. ; The average grayscale reference acquisition module is used to calculate the average value by traversing the compensated red channel pixel values. After traversing and correcting the green channel pixels Calculate the average value Traversing the corrected blue channel pixel values Calculate the average value ,according to , and The average grayscale reference of the three channels was calculated. ; The second pixel correction module is used to correct the pixel values ​​of the pixels. Each pixel of the channel is multiplied / Obtain the corrected R pixel value; for the Each pixel of the channel is multiplied / Obtain the corrected G pixel value; for the Each pixel of the channel is multiplied / Obtain the corrected B pixel value; The combination module is used to combine the corrected R, G, and B pixel values ​​according to their corresponding pixel coordinates to obtain an enhanced image.

9. A personnel platform management method and system based on electronic fences according to claim 6, characterized in that, The target data acquisition module includes improvements based on YOLOv11, wherein: In the YOLOv11 backbone network, the YOLOv11 Layer 2 C3K2 module, Layer 4 C3K2 module, and Layer 6 C3K2 module are replaced with ZAWRCNet modules; In the YOLOv11 backbone network, the YOLOv11 layer 7 Conv and layer 8 C3K2 modules are removed; The ZAWRCNet module works as follows: Obtain the input features, input the input features into the Conv module to obtain the first convolutional feature, and split the first convolutional feature to obtain the first channel feature and the second channel feature; The first channel feature is input into the Star Res-CAA Block module to obtain the first initial feature; the first initial feature is input into the Star Res-CAA Block module to obtain the second initial feature; The Star Res-CAA Block module comprises a dimensionality reduction layer, a feature extraction layer, and a dimensionality increase layer in sequence. The first convolutional feature is input into the Conv module to obtain the second convolutional feature; The first convolutional feature is sequentially input into the Conv module, DWConv module, and Point-wise Conv module to obtain the third convolutional feature; The first target feature is obtained by concatenating the first channel feature, the second channel feature, the first initial feature, the second convolutional feature, the second initial feature, and the third convolutional feature. The first target feature is then input into the Conv module to obtain the second target feature. The second target feature is used as the final output image of the ZAWRCNet module.

10. A personnel platform management method and system based on electronic fences according to claim 6, characterized in that, The task allocation module includes a distance calculation module, a priority module, and a dynamic adjustment module, wherein: The distance calculation module is used to calculate the distance data between the real-time location of each security personnel and the accident location based on the target data and the working status data. The priority module is used to calculate the priority value of security personnel relative to other personnel based on the work status data and the distance data, and adjust the task allocation order and task content of security personnel in the security instructions according to the priority value; The dynamic adjustment module is used to track the dynamic location of security personnel in real time, update their distance data, and dynamically adjust task allocation.

Images