Cable well lid intelligent monitoring method based on image processing

Abstract

A cable well lid intelligent monitoring method based on image processing belongs to the technical field of image processing, and comprises the following steps: step 1, constructing a hardware system for cable well lid intelligent monitoring; 2, when the well lid of the cable well is opened, the central processing module triggers monitoring; 3, after monitoring is triggered, the central processing module carries out operation target recognition and positioning; 4, after the operation target is recognized and positioned, the central processing module analyzes and records operation behaviors; 5, after the operation behaviors are analyzed and recorded, the central processing module executes operation planning judgment; and 6, after the operation planning judgment, the central processing module carries out data backtracking and analysis. According to the cable well operation and maintenance method, accurate recognition of underground operation personnel and operation tools is achieved through the image processing technology, operation key information is automatically recorded, the operation legality is judged in combination with a planned operation database, and then the safety, normalization and intelligent level of cable well operation and maintenance are improved.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology, specifically relating to an intelligent monitoring method for cable well covers based on image processing. Background Technology

[0002] In urban power and communication network systems, cable wells are critical infrastructure, making the safe management of their manhole covers and the standardized monitoring of downhole operations crucial. Existing cable well covers with temperature detection functions, such as the one described in patent publication number "CN221502067U," while improving anti-theft performance through fixing mechanisms and monitoring environmental and internal well temperatures using temperature detectors, still have significant technical shortcomings in actual operation and maintenance: they cannot identify, record, or plan downhole operations.

[0003] Currently, well-entry operations largely rely on manual reporting and on-site supervision, lacking automated monitoring methods. On the one hand, unauthorized personnel may enter the well by damaging the manhole cover or breaking through simple protective devices, causing safety accidents such as cable theft and equipment damage, and it is difficult to trace the perpetrators and the process afterward. On the other hand, key information such as the start and end times, personnel identities, and operation duration of legitimate maintenance operations must be manually recorded, which is prone to omissions and errors, leading to non-standard operation procedures and difficulty in defining responsibilities. In addition, existing technology cannot determine in advance whether well-entry operations are planned maintenance tasks, making it difficult to distinguish between legitimate operations and illegal intrusions, posing a significant challenge to the intelligent management of cable wells.

[0004] With the rapid development of image processing technology, target detection and behavior analysis technologies based on cameras and deep learning algorithms have been widely used in security, transportation, and other fields, enabling accurate identification and behavior trajectory tracking of specific targets (such as human bodies and tools). Combining this technology with cable well cover systems can effectively compensate for the shortcomings of existing technologies in monitoring operational behavior, constructing a comprehensive management solution integrating anti-theft, temperature monitoring, intelligent identification and recording of operational behavior, and improving the intelligence and safety of cable well operation and maintenance. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention proposes an intelligent monitoring method for cable well covers based on image processing. This method aims to overcome the limitations of existing cable well covers in identifying and recording downhole operations and determining the planning nature of these operations. Through image processing technology, it achieves accurate identification of downhole personnel and tools, automatically records key operational information, and combines this with a planned operation database to determine the legality of the operation, thereby improving the safety, standardization, and intelligence level of cable well maintenance.

[0006] The present invention employs the following technical solution.

[0007] A method for intelligent monitoring of cable well covers based on image processing includes: Step 1: Construct a hardware system for intelligent monitoring of cable well covers; Step 2: When the manhole cover of the cable well is opened, the central processing module triggers monitoring; Step 3: After monitoring is triggered, the central processing module identifies and locates the work target; Step 4: After the target of the operation is identified and located, the central processing module analyzes and records the operation behavior; Step 5: After analyzing and recording the work behavior, the central processing module performs a work planning judgment; Step 6: After the work plan is determined, the central processing module performs data backtracking and analysis.

[0008] Furthermore, in step 1, the hardware system includes a manhole cover body with temperature detection function, a mounting base, a temperature detector, a fixing mechanism, a picking mechanism, and a sensor, and also adds the following components: Image acquisition module: includes a high-definition infrared camera installed on the lower surface of the manhole cover body and a panoramic camera installed on the inner side of the manhole wall. The high-definition infrared camera and the panoramic camera are used to acquire image and video data of the area inside the manhole at the moment the manhole cover is opened. Data storage module: Built-in SD card or external cloud storage server. The data storage module is used to store image and video data, job record data and planned job database acquired by the image acquisition module. Data transmission module: Using 4G, 5G or LoRa modules, it enables real-time transmission of image data, recognition results and work records to a remote monitoring platform that is connected to the data transmission module; The central processing module, which is connected to the image acquisition module, data storage module, data transmission module, and sensors, is equipped with an embedded processor and runs target detection algorithms, behavior analysis algorithms, and data comparison programs to achieve image data processing, work behavior recognition, and planning judgment.

[0009] Furthermore, step 2 specifically includes: When the sensor triggers the fixing mechanism to unlock via the sensor card, the central processing module receives the unlock signal and activates the image acquisition module. This module controls the high-definition infrared camera on the lower surface of the manhole cover to capture images of the opened area of ​​the manhole cover and transmit them to the central processing module. Simultaneously, the panoramic camera on the manhole wall is activated to start recording video inside the manhole and transmit it to the central processing module.

[0010] Furthermore, step 3 specifically includes: The central processing module inputs the image and video data acquired by the image acquisition module into the pre-trained improved YOLOv8 target detection model to identify and locate targets in the image. The identified targets include human bodies and tools.

[0011] Furthermore, in step 3, the specific process of identifying and locating the target in the image is as follows: Step 3-1: Preprocess the input image by enhancing image contrast through adaptive histogram equalization; Step 3-2: Input the preprocessed image into the improved YOLOv8 object detection model for further processing.

[0012] Furthermore, step 3-2 specifically includes: The improved YOLOv8 object detection model extracts image features through a backbone network, achieves multi-scale feature fusion using a neck network, and finally outputs the bounding box coordinates of the target through a head network. , , , ), target category (Class) and confidence level (Confidence), among which The x-coordinate of the center of the target bounding box. The x-coordinate of the center of the target bounding box; and These represent the width and height of the target bounding box, respectively; Class is the target category label. Set confidence threshold When the confidence level of a target exceeds a confidence threshold, it is considered a valid target, and its bounding box coordinates and category information are retained; otherwise, they are not retained. Its bounding box coordinates and category information.

[0013] Furthermore, in step 3-2, it is determined to be a valid target. The formula is as follows: ; Where Class∈{“Human Body”,“Work Tools”}.

[0014] Furthermore, step 4 specifically includes: Step 4-1: The central processing module performs worker counting and identity association, that is, it counts the number of workers inside the well by using the "human body" category bounding box in the target detection results. Meanwhile, if the camera captures the QR code or RFID tag on the safety helmet worn by the worker, the tag information is extracted through image recognition algorithm and compared with the maintenance personnel database of the remote monitoring platform to obtain the name and employee ID information of the worker that matches the tag information. Step 4-2: The central processing module performs tool tracking and operation time recording. Based on the bounding box of the "tool" in the target detection result, the Kalman filter algorithm is used to track the movement trajectory of the tool, determine whether the tool is used for cable maintenance, and record the opening time T1 and closing time T2 of the manhole cover. The operation time ΔT = T2 - T1. Step 4-3: The central processing module generates a work record, which involves associating and storing the worker's identity information, number of workers N, type of work tools, work duration ΔT, and key video clips during the work into the data storage module to generate a structured work record.

[0015] Furthermore, step 5 specifically includes: The central processing module retrieves the planned operation database of the remote monitoring platform through the data transmission module. This database stores information on planned maintenance tasks, including the planned operation time interval, the employee number of the planned operator, and the planned operation area. The module compares the cable well number, operator employee number, and operation time T1 in the current operation record with the information in the database using the following formula: Wherein, Legitimacy is the result of the job legality judgment, with a value of "Legal (planned job)" or "Illegal (illegal intrusion)", used to distinguish between legitimate operation and maintenance and illegal behavior; T_start is the start time of the planned job, in the format "YYYY-MM-DDHH:MM:SS", from the planned job database; T_end is the end time of the planned job, in the format "YYYY-MM-DDHH:MM:SS", from the planned job database; the cable well number is a unique identifier for each cable well; the employee number is a unique identifier for the operation and maintenance personnel; If an illegal intrusion is detected, the central processing module immediately sends an alarm signal to the remote monitoring platform through the data transmission module. At the same time, it triggers the audible and visual alarm on the manhole cover that is connected to the central processing module, emitting a piercing alarm sound and flashing red light to deter unauthorized personnel.

[0016] Furthermore, step 6 specifically includes: The remote monitoring platform can view work records and video data in real time, and can also search historical work records by keywords such as time, cable well number, and worker's employee number, for use in operation and maintenance process auditing, accident tracing, and work efficiency analysis.

[0017] The beneficial effects of the present invention are as follows: Compared with the prior art, the technical effects of the present invention include: Achieve full-process monitoring of underground operations: Through image acquisition modules and improved YOLOv8 models, accurately identify personnel and tools, and automatically record the number of personnel, their identities, duration, and tool usage. This solves the shortcomings of existing technologies that cannot record operational behavior, enables traceability of the operation process, and avoids omissions and errors in manual recording.

[0018] Accurately distinguish between legal operations and illegal intrusions: By combining a planned operation database with multi-dimensional information comparison, the legality of entering the well can be quickly determined. In case of illegal intrusion, an alarm is triggered in real time, effectively deterring unauthorized personnel, reducing the risk of cable theft and equipment damage, and improving the anti-theft security of cable wells.

[0019] Enhance the intelligence level of operation and maintenance management: The remote monitoring platform enables real-time viewing and historical review of operation data, supports operation and maintenance process auditing and operation efficiency analysis, helps managers optimize operation and maintenance plans, reduces manual supervision costs, and promotes the transformation of cable well operation and maintenance from "manual management" to "intelligent monitoring".

[0020] Adaptable to complex working environments: The combination of high-definition infrared cameras and 360° panoramic cameras ensures monitoring effectiveness in different lighting conditions such as daytime, nighttime, and low light, with no blind spots; the improved YOLOv8 model has a high recognition accuracy, ensuring the reliability of target detection and meeting the diverse needs of actual working scenarios. Attached Figure Description

[0021] Figure 1 This is a flowchart of the intelligent monitoring method for cable well covers based on image processing described in this invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, any other embodiments obtained by those skilled in the art without creative effort are all within the protection scope of this invention.

[0023] like Figure 1 As shown, the intelligent monitoring method for cable well covers based on image processing according to the present invention includes: Step 1: Construct a hardware system for intelligent monitoring of cable well covers; In a preferred but non-limiting embodiment of the present invention, in step 1, the present invention is implemented based on a hardware system of a cable well cover with temperature detection function. The hardware system includes a well cover body with temperature detection function, a mounting base, a temperature detector, a fixing mechanism, a picking mechanism, and a sensor, and also adds the following components: Image acquisition module: includes a high-definition infrared camera (supporting shooting in night and low light environments) installed on the lower surface of the manhole cover body and a 360° panoramic camera installed on the inner side of the manhole wall. The high-definition infrared camera and the panoramic camera are used to acquire image and video data of the area inside the manhole at the moment the manhole cover is opened. Data storage module: Built-in SD card or external cloud storage server. The data storage module is used to store image and video data, job record data and planned job database acquired by the image acquisition module. Data transmission module: Employs a 4G, 5G, or LoRa wireless communication module to enable real-time transmission of image data, recognition results, and work records to a remote monitoring platform that is connected to the data transmission module; the remote monitoring platform can be a computer or a server.

[0024] The central processing module, which is connected to the image acquisition module, data storage module, data transmission module, and sensors, is equipped with an embedded processor (such as the ARM Cortex-A series) and runs target detection algorithms, behavior analysis algorithms, and data comparison programs to realize image data processing, job behavior recognition, and planning judgment.

[0025] This intelligent monitoring method specifically includes the following steps: Step 2: When the manhole cover of the cable well is opened, the central processing module triggers monitoring; In a preferred but non-limiting embodiment of the present invention, step 2 specifically includes: When the sensor triggers the fixing mechanism to unlock via the sensor card (the motor drives the threaded rod to rotate, and the plug block disengages from the mounting hole), the central processing module receives the unlock signal and immediately starts the image acquisition module. It controls the high-definition infrared camera on the lower surface of the manhole cover to capture images of the opened area of ​​the manhole cover and transmits them to the central processing module. At the same time, the panoramic camera on the manhole wall starts recording video inside the manhole and transmits it to the central processing module. The acquisition frequency is set to 25 frames per second to ensure no video lag.

[0026] Step 3: After monitoring is triggered, the central processing module identifies and locates the work target; In a preferred but non-limiting embodiment of the present invention, step 3 specifically includes: The central processing module inputs the image and video data acquired by the image acquisition module into the pre-trained improved YOLOv8 target detection model to identify and locate targets in the image. The identified targets include human bodies and tools (such as wrenches, screwdrivers, and multimeters).

[0027] In a preferred but non-limiting embodiment of the present invention, the specific process of identifying and locating the target in the image in step 3 is as follows: Step 3-1: Preprocess the input image by using Adaptive Histogram Equalization (CLAHE) to enhance image contrast and remove noise interference in low-light environments; Step 3-2: Input the preprocessed image into the improved YOLOv8 object detection model for further processing.

[0028] In a preferred but non-limiting embodiment of the present invention, step 3-2 specifically includes: The improved YOLOv8 object detection model extracts image features through a backbone network (CSPDarknet53), achieves multi-scale feature fusion using a neck network (FPN+PAN), and finally outputs the bounding box coordinates of the target through a head network. , , , ), target category (Class) and confidence level (Confidence), among which The x-coordinate of the center of the target bounding box. The x-coordinate of the center of the target bounding box. and The unit is pixels (px), used to locate the position of the target in the image; and These are the width and height of the target bounding box, respectively, in pixels (px), used to describe the size of the target; Class is the target category label, with values ​​of "human body" or "work tool", used to distinguish the core target in the work scene; Confidence is the confidence of the improved YOLOv8 target detection model in the target category, with a value range of [0,1], representing the reliability of the improved YOLOv8 target detection model in determining the target to be of the corresponding category; Set confidence threshold ( (≥0.8) When the confidence score of a target exceeds the confidence threshold, it is considered a valid target, and its bounding box coordinates and category information are retained; otherwise, they are not retained. Its bounding box coordinates and category information. The confidence threshold is defined as [0,1], and is set in this method. =0.8, used to filter out false positives in target detection. When the confidence score of the improved YOLOv8 target detection is ≥0.8, it is judged as a real target, avoiding misjudgment caused by low confidence scores.

[0029] In a preferred but non-limiting embodiment of the present invention, in step 3-2, the target is determined to be a valid target. The formula is as follows: ; Where Class∈{“Human Body”,“Work Tools”}.

[0030] Step 4: After the target of the operation is identified and located, the central processing module analyzes and records the operation behavior; In a preferred but non-limiting embodiment of the present invention, step 4 specifically includes: Step 4-1: The central processing module performs worker counting and identity association, that is, it counts the number of workers inside the well by using the "human body" category bounding box in the target detection results. Meanwhile, if the camera captures the QR code or RFID tag on the safety helmet worn by the worker, the tag information is extracted through image recognition algorithm and compared with the maintenance personnel database of the remote monitoring platform to obtain the worker's name, employee number and other identity information that match the tag information. Step 4-2: The central processing module performs tool tracking and operation time recording. Based on the bounding box of the "tool" in the target detection result, the Kalman filter algorithm is used to track the movement trajectory of the tool and determine whether the tool is used for cable maintenance (such as when the tool is close to the cable joint area). At the same time, the opening time T1 and closing time T2 (the time when the sensor triggers the locking mechanism) of the manhole cover are recorded. The operation time ΔT = T2 - T1. ΔT is the duration of the operation, expressed in hours (h) or minutes (min). It reflects the duration of the operation and is used to assess operation efficiency and determine whether there is any overtime. T1 represents the opening time of the manhole cover, in the format "YYYY-MM-DDHH:MM:SS", which is the time when the sensor triggers the locking mechanism to unlock. T2 represents the manhole cover closing time, in the format "YYYY-MM-DDHH:MM:SS", which is the time when the sensor triggers the locking mechanism.

[0031] Step 4-3: The central processing module generates work records, which involves associating and storing the identity information of the workers, the number of workers N, the type of work tools, the work duration ΔT, and key video clips during the work (such as the moment of tool operation and the moment of personnel entering and leaving the well) to the data storage module to generate structured work records.

[0032] The format of structured job logs includes the following: Work record ID, worker's name, employee number, number of workers The task tool type, start time T1, stop time T2, task duration ΔT, and key video path.

[0033] Work record ID, worker's name, employee number, number of workers The job tool type, start time T1, stop time T2, job duration ΔT, and key video path are 20240927001, Zhang San, 001, 2, wrench and multimeter, 2024-09-27 09:00:00, 2024-09-27 11:30:00, 2.5h, and / storage / video / 20240927001.mp4, respectively.

[0034] Step 5: After analyzing and recording the work behavior, the central processing module performs a work planning judgment; In a preferred but non-limiting embodiment of the present invention, step 5 specifically includes: The central processing module retrieves the planned operation database of the remote monitoring platform through the data transmission module. This database stores information on planned maintenance tasks, including the planned operation time interval [T_start, T_end], the planned operator's employee number, and the planned operation area (cable well number). The current operation record's cable well number, operator's employee number, and operation time T1 are compared with the information in the database using the following formula: Wherein, Legitimacy is the result of the job legality judgment, with a value of "Legal (planned job)" or "Illegal (illegal intrusion)", used to distinguish between legitimate operation and maintenance and illegal behavior; T_start is the start time of the planned job, in the format "YYYY-MM-DDHH:MM:SS", from the planned job database; T_end is the end time of the planned job, in the format "YYYY-MM-DDHH:MM:SS", from the planned job database; the cable well number is a unique identifier for each cable well (e.g., "JD-001"), used to determine the specific location of the current operation; the worker number is the unique identifier of the operation and maintenance personnel (e.g., "001"), used to verify whether the operator is a planned personnel; If an unauthorized intrusion is detected, the central processing module immediately sends an alarm signal to the remote monitoring platform through the data transmission module. At the same time, it triggers the audible and visual alarm (a newly added component) on the manhole cover that is connected to the central processing module, emitting a piercing alarm sound and flashing red light to deter unauthorized personnel.

[0035] Step 6: After the work plan is determined, the central processing module performs data backtracking and analysis.

[0036] In a preferred but non-limiting embodiment of the present invention, step 6 specifically includes: The remote monitoring platform can view work records and video data in real time, and can also search historical work records by keywords such as time, cable well number, and worker's employee number, which can be used for operation and maintenance process auditing, accident tracing, and work efficiency analysis.

[0037] The following is a specific example of the present invention: (a) Hardware deployment Image acquisition module installation: A 5cm diameter mounting hole is made at the center of the lower surface of the manhole cover to embed a high-definition infrared camera (model: Hikvision DS-2CD3T46WD-I3, 4-megapixel resolution, supports 0.001Lux ultra-low illumination). The camera lens is facing downwards to cover the manhole area after the manhole cover is opened, and the viewing angle is adjusted to 90° to ensure that the complete image of people entering and exiting the manhole is captured. A 360° panoramic camera (model: Dahua DH-SDT5043E, 8-megapixel resolution, supports fisheye correction) is fixed to the inside of the cable well wall at a distance of 1.5m from the well opening using expansion bolts, with the lens facing upwards to cover the entire well space and ensure no blind spots in monitoring.

[0038] Integration with other modules: The central processing module uses a Raspberry Pi 4B embedded development board (equipped with an ARM Cortex-A72 processor and 4GB of memory), which is installed in a waterproof box inside the manhole cover body. It is connected to the temperature detector, motor, sensor, and image acquisition module through the GPIO interface to realize signal interaction. The data storage module uses a 128GB industrial-grade SD card (supports continuous writing and is resistant to high and low temperatures), which is inserted into the SD card slot of the Raspberry Pi 4B. At the same time, it establishes a connection with the cloud storage server (Alibaba Cloud OSS) through a 4G module (model: Huawei ME909u-521) to achieve dual local and cloud storage. The audible and visual alarm (model: LTE-1101J) is installed on the upper edge of the manhole cover body and connected to the central processing module via a wire. When triggered, it emits a 110dB alarm sound and flashes a red LED light.

[0039] (II) Software Implementation Improved YOLOv8 model training: A dataset of cable well operation scenarios was constructed, containing 5,000 images covering different lighting conditions (daytime and nighttime), different numbers of workers (1-3 people), and different tools (wrench, screwdriver, multimeter, etc.). The "human body" and "tool" in the images were labeled in YOLO format (category ID + center coordinate x / width + center coordinate y / height + width / width + height / height). The transfer learning method was adopted, based on the pre-trained YOLOv8-s model, and fine-tuned on the constructed dataset. The training environment was PyTorch2.0, the learning rate was set to 0.001, the batch size was 16, and the number of iterations was 100. The model parameters were adjusted through cross-validation. The final model achieved an accuracy of 98.5% in "human body" recognition and 96.2% in "homework tool" recognition on the test set, which meets the needs of practical applications.

[0040] Central Processing Unit (CPU) Program Development: The program was developed using Python, and the OpenCV library was used to read the video stream from the image acquisition module. The trained YOLOv8 model was then used for object detection. Use the pyserial library to receive unlock / lock signals from the sensors, triggering the camera to start and stop. The requests library is used to implement HTTP communication with the remote monitoring platform to send job records and alarm signals. The SQLite database is used to store job records locally, while the video data is uploaded to cloud storage via Alibaba Cloud SDK.

[0041] Remote monitoring platform setup: A remote monitoring platform was built based on the Django web framework. The front end used Vue.js to implement data visualization, and the back end used a MySQL database to store planned job data and historical job records. The platform's functions include: real-time video viewing, work record retrieval (by time, cable well number, and work number), planned work task management (adding, editing, and deleting tasks), and alarm information push (SMS and platform pop-ups). Managers can log in to the platform via computer or mobile phone to achieve remote monitoring.

[0042] (III) Practical Application Process Preparation for planned operations: Maintenance personnel create planned operation tasks in advance on the remote monitoring platform, and enter information such as "cable well number (e.g., JD-001)," "planned operation time range (2024-09-27 09:00:00-2024-09-27 12:00:00)," and "operator employee number (001, 002)." The platform stores the task data in the planned operation database.

[0043] On-site operation execution: When maintenance personnel arrive at the site, they place the sensor card close to the sensor on the manhole cover. The sensor triggers the fixing mechanism to unlock (the motor drives the threaded rod to rotate, and the plug block disengages from the mounting hole). The central processing module receives the unlock signal and immediately activates the high-definition infrared camera and the panoramic camera. The camera captures images of the area where the manhole cover is opened. The central processing module uses an improved YOLOv8 model to identify two "human bodies" and two types of tools: a "wrench" and a "multimeter." It also identifies the worker numbers "001" and "002" on the workers' safety helmets and matches them with the worker numbers in the planned operation database. During the operation, the panoramic camera continuously recorded video, and the Kalman filter algorithm tracked the movement trajectory of the tool. The record showed that the tool was always operating in the cable joint area, which was determined to be normal maintenance work. After the operation is completed, the maintenance personnel trigger the locking of the fixed mechanism through the proximity card. The central processing module records the closing time T2=2024-09-27 11:30:00, calculates the operation duration ΔT=2.5h, generates the operation record and uploads it to the remote monitoring platform. The platform determines the operation as "Legal (planned operation)".

[0044] Handling of unauthorized intrusion: If unauthorized personnel forcefully damage the manhole cover to trigger the unlocking of the fixing mechanism, the central processing module will activate the camera, identify the "human body" but not find a matching planned task, and immediately determine it as "Illegal (unauthorized intrusion)", send an alarm signal to the remote monitoring platform, and trigger the audible and visual alarm. After receiving the alarm, the management personnel can view the real-time video through the platform and arrange personnel to go to the scene to handle the situation.

[0045] The beneficial effects of the present invention are as follows: Compared with the prior art, the technical effects of the present invention include: Achieve full-process monitoring of underground operations: Through image acquisition modules and improved YOLOv8 models, accurately identify personnel and tools, and automatically record the number of personnel, their identities, duration, and tool usage. This solves the shortcomings of existing technologies that cannot record operational behavior, enables traceability of the operation process, and avoids omissions and errors in manual recording.

[0046] Accurately distinguish between legal operations and illegal intrusions: By combining a planned operation database with multi-dimensional information comparison, the legality of entering the well can be quickly determined. In case of illegal intrusion, an alarm is triggered in real time, effectively deterring unauthorized personnel, reducing the risk of cable theft and equipment damage, and improving the anti-theft security of cable wells.

[0047] Enhance the intelligence level of operation and maintenance management: The remote monitoring platform enables real-time viewing and historical review of operation data, supports operation and maintenance process auditing and operation efficiency analysis, helps managers optimize operation and maintenance plans, reduces manual supervision costs, and promotes the transformation of cable well operation and maintenance from "manual management" to "intelligent monitoring".

[0048] Adaptable to complex working environments: The combination of high-definition infrared cameras and 360° panoramic cameras ensures monitoring effectiveness in different lighting conditions such as daytime, nighttime, and low light, with no blind spots; the improved YOLOv8 model has a high recognition accuracy, ensuring the reliability of target detection and meeting the diverse needs of actual working scenarios.

[0049] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention without departing from the spirit and scope of the present invention. Any modifications or equivalent substitutions should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for intelligent monitoring of cable well covers based on image processing, characterized in that, include: Step 1: Construct a hardware system for intelligent monitoring of cable well covers; Step 2: When the manhole cover of the cable well is opened, the central processing module triggers monitoring; Step 3: After monitoring is triggered, the central processing module identifies and locates the work target; Step 4: After the target of the operation is identified and located, the central processing module analyzes and records the operation behavior; Step 5: After analyzing and recording the work behavior, the central processing module performs a work planning judgment; Step 6: After the work plan is assessed, the central processing module performs data backtracking and analysis; In step 1, the hardware system includes a manhole cover body with temperature detection function, a mounting base, a temperature detector, a fixing mechanism, a picking mechanism, and a sensor, and also adds the following components: Image acquisition module: includes a high-definition infrared camera installed on the lower surface of the manhole cover body and a panoramic camera installed on the inner side of the manhole wall. The high-definition infrared camera and the panoramic camera are used to acquire image and video data of the area inside the manhole at the moment the manhole cover is opened. Data storage module: Built-in SD card or external cloud storage server. The data storage module is used to store image and video data, job record data and planned job database acquired by the image acquisition module. Data transmission module: Using 4G, 5G or LoRa modules, it enables real-time transmission of image data, recognition results and work records to a remote monitoring platform that is connected to the data transmission module; The central processing module, which is connected to the image acquisition module, data storage module, data transmission module, and sensors, is equipped with an embedded processor and runs target detection algorithms, behavior analysis algorithms, and data comparison programs to achieve image data processing, work behavior recognition, and planning judgment.

2. The intelligent monitoring method for cable well covers based on image processing according to claim 1, characterized in that, Step 2 specifically includes: When the sensor triggers the fixing mechanism to unlock via the sensor card, the central processing module receives the unlock signal and activates the image acquisition module. This module controls the high-definition infrared camera on the lower surface of the manhole cover to capture images of the opened area of ​​the manhole cover and transmit them to the central processing module. Simultaneously, the panoramic camera on the manhole wall is activated to start recording video inside the manhole and transmit it to the central processing module.

3. The intelligent monitoring method for cable well covers based on image processing according to claim 2, characterized in that, Step 3 specifically includes: The central processing module inputs the image and video data acquired by the image acquisition module into the pre-trained improved YOLOv8 target detection model to identify and locate targets in the image. The identified targets include human bodies and tools.

4. The intelligent monitoring method for cable well covers based on image processing according to claim 3, characterized in that, In step 3, the specific process of identifying and locating targets in the image is as follows: Step 3-1: Preprocess the input image by enhancing image contrast through adaptive histogram equalization; Step 3-2: Input the preprocessed image into the improved YOLOv8 object detection model for further processing.

5. The intelligent monitoring method for cable well covers based on image processing according to claim 4, characterized in that, Step 3-2 specifically includes: The improved YOLOv8 object detection model extracts image features through a backbone network, achieves multi-scale feature fusion using a neck network, and finally outputs the bounding box coordinates of the target through a head network. , , , ), target category (Class) and confidence level (Confidence), among which The x-coordinate of the center of the target bounding box. The x-coordinate of the center of the target bounding box; and These represent the width and height of the target bounding box, respectively; Class is the target category label. Set confidence threshold When the confidence level of a target exceeds a confidence threshold, it is considered a valid target, and its bounding box coordinates and category information are retained; otherwise, they are not retained. Its bounding box coordinates and category information.

6. The intelligent monitoring method for cable well covers based on image processing according to claim 5, characterized in that, In step 3-2, it is determined to be a valid target. The formula is as follows: ； Where Class∈{"Human Body","Work Tools"}.

7. The intelligent monitoring method for cable well covers based on image processing according to claim 6, characterized in that, Step 4 specifically includes: Step 4-1: The central processing module performs worker counting and identity association, that is, it counts the number of workers inside the well by using the "human body" category bounding box in the target detection results. Meanwhile, if the camera captures the QR code or RFID tag on the safety helmet worn by the worker, the tag information is extracted through image recognition algorithm and compared with the maintenance personnel database of the remote monitoring platform to obtain the name and employee ID information of the worker that matches the tag information. Step 4-2: The central processing module performs tool tracking and operation time recording. Based on the bounding box of the "tool" in the target detection result, the Kalman filter algorithm is used to track the movement trajectory of the tool, determine whether the tool is used for cable maintenance, and record the opening time T1 and closing time T2 of the manhole cover. The operation time ΔT = T2 - T1. Step 4-3: The central processing module generates a work record, which involves associating and storing the worker's identity information, number of workers N, type of work tools, work duration ΔT, and key video clips during the work into the data storage module to generate a structured work record.

8. The intelligent monitoring method for cable well covers based on image processing according to claim 7, characterized in that, Step 5 specifically includes: The central processing module retrieves the planned operation database of the remote monitoring platform through the data transmission module. This database stores information on planned maintenance tasks, including the planned operation time interval, the employee number of the planned operator, and the planned operation area. The module compares the cable well number, operator employee number, and operation time T1 in the current operation record with the information in the database using the following formula: Wherein, Legitimacy is the result of the job legality judgment, with a value of "Legal" or "Illegal", used to distinguish between legal operation and maintenance and illegal behavior; T_start is the start time of the planned job, in the format "YYYY-MM-DDHH:MM:SS", from the planned job database; T_end is the end time of the planned job, in the format "YYYY-MM-DDHH:MM:SS", from the planned job database; the cable well number is a unique identifier for each cable well; the employee number is a unique identifier for the operation and maintenance personnel; If an illegal intrusion is detected, the central processing module immediately sends an alarm signal to the remote monitoring platform through the data transmission module. At the same time, it triggers the audible and visual alarm on the manhole cover that is connected to the central processing module, emitting a piercing alarm sound and flashing red light to deter unauthorized personnel.

9. The intelligent monitoring method for cable well covers based on image processing according to claim 8, characterized in that, Step 6 specifically includes: The remote monitoring platform can view work records and video data in real time, and can also search historical work records by keywords such as time, cable well number, and worker's employee number, for use in operation and maintenance process auditing, accident tracing, and work efficiency analysis.

Images