A camera and solar-based cable channel anti-external damage AI intelligent monitoring method and system

Abstract

The application relates to the technical field of power cable channel safety monitoring, and discloses an AI intelligent monitoring method and system for cable channel external damage prevention based on a camera and solar energy, which can calculate the sustainable working time and automatically switch the working mode by monitoring the residual capacity of the energy storage battery and the solar power generation power in real time, dynamically adjust the image acquisition frame rate according to the risk level, demarcate the key monitoring area for target detection, identify the external damage risk targets such as excavators and cranes, calculate the actual distance between the target and the cable channel, determine the risk in combination with the motion trend, trigger the sound-light alarm and upload the background management platform; the application realizes intelligent switching of the working mode through dynamic energy efficiency management, realizes local low-delay recognition through deployment of the edge-end lightweight model, realizes fine energy consumption management and control through the adaptive frame rate strategy driven by the risk level, realizes accurate alarm pushing through construction record linkage, prolongs the endurance time, and reduces the false alarm rate and the missed alarm rate.

Description

Technical Field

[0001] This invention relates to the field of power cable channel safety monitoring technology, and more specifically to an AI-powered intelligent monitoring method and system for preventing external damage to cable channels based on cameras and solar energy. Background Technology

[0002] With the continuous expansion of urban power grids, power cable lines have become an important part of urban distribution networks. Most cables are laid underground in the form of direct burial or conduits. However, in urban road reconstruction, municipal construction, rail transit construction and other activities, large construction machinery such as excavators, cranes, and pile drivers are very likely to cause external damage to underground cable channels, resulting in frequent line tripping and power outage accidents, which seriously threaten the safe operation of the power grid and public safety. According to incomplete statistics, cable faults caused by external damage account for more than 40% of urban power distribution network failures, and this percentage is increasing year by year. Therefore, real-time monitoring and timely early warning of construction machinery around cable channels are of great significance for ensuring the safe operation of the power grid. Currently, the monitoring of cable channels against external damage still faces the following problems: Cable channels are mostly distributed along urban roadsides and in suburban areas lacking mains power interfaces. Existing solar power supply solutions are greatly affected by weather and seasons, resulting in intermittent and unstable power generation. Physical sensing methods cannot identify the specific type and precise location of external damage sources. Cloud-based recognition architectures require transmitting images to the cloud for processing, leading to high latency, large bandwidth consumption, and loss of recognition capabilities for front-end devices when the network is unstable. Existing equipment lacks refined power management strategies. Cameras often operate continuously at full frame rate, adjusting only based on remaining battery power without considering real-time solar power generation, making it impossible to predict future energy replenishment. In extreme cases, they are prone to shutdown due to battery depletion. Fixed parameters are used for acquisition frame rate and recognition thresholds, which cannot be dynamically adjusted according to lighting conditions, risk levels, and time of day. There is a lack of construction registration verification mechanisms, making it impossible to distinguish between registered compliant construction and illegal destructive construction. Invalid alarms reduce the attention paid to genuine alarms by maintenance personnel. Summary of the Invention

[0003] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide an AI intelligent monitoring method and system for preventing external damage to cable channels based on cameras and solar energy, so as to solve the technical problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, comprising the following steps: Step S1: Deploy monitoring devices above or to the side of the cable channel, and configure monitoring areas, warning areas, and alarm parameters; Step S2: The solar photovoltaic panel converts light energy into electrical energy and stores it in the energy storage battery. The remaining power of the energy storage battery and the power generation of the solar photovoltaic panel are monitored in real time. Step S3: Calculate the sustainable working time CX and adjust the mode accordingly; Step S4: Calculate the current image acquisition frame rate PL; Step S5: Based on the cable channel's direction and width, as well as the camera's current installation posture, dynamically delineate key monitoring areas in the acquired video images, and only perform target detection and recognition processing on images within the key monitoring areas; Step S6: Perform real-time analysis on video images within the key monitoring area, and call the lightweight target detection model to identify external damage risk targets in the images. The external damage risk targets include excavators, cranes, and pile drivers. Step S7: When an external damage risk target is identified, calculate the actual distance D between the target and the cable channel; Step S8: Continuously track the target entering the warning area, calculate the target's velocity vector and direction of movement, and determine whether the target's direction of movement points towards the center line of the cable channel. When the target enters the warning area and its movement trend points towards the cable channel, it is determined that there is a risk of external damage. Step S9: After determining that there is a risk of external damage, issue an on-site audible and visual alarm and upload the alarm information to the back-end management platform.

[0005] In a preferred embodiment, the formula for calculating the sustainable working time CX in step S3 is as follows: In the formula, SY represents the current remaining capacity of the energy storage battery, GP represents the current measured power generation of the solar photovoltaic panel, and XL represents the charging conversion efficiency. YL represents the preset prediction time window, YL represents the reserved safe power, and GH represents the system power consumption in the current operating mode.

[0006] In a preferred embodiment, in step S3, the calculated continuous working time CX is compared with a preset first threshold Y1 and a second threshold Y2, wherein the first threshold Y1 > the second threshold Y2. When the continuous working time CX > the first threshold Y1, the normal working mode is maintained and the camera continuously acquires images. When the first threshold Y1 ≥ the continuous working time CX > the second threshold Y2, the mode is switched to low power mode and the camera switches to interval acquisition. When the second threshold Y2 ≥ the continuous working time, the mode is switched to sleep mode and the camera stops working, retaining only the wireless communication function.

[0007] In a preferred embodiment, in step S4, the formula for calculating the current image acquisition frame rate PL is: In the formula The minimum capture frame rate, The highest capture frame rate, This represents the current risk level value for the monitored area. At the highest risk level, the camera captures video images of the cable channel and surrounding area according to the calculated current image acquisition frame rate PL.

[0008] In a preferred embodiment, the risk level value assessment method in step S4 is as follows: time period factor accounts for 40% of the weight, based on whether the current period is in a preset high-incidence construction period; seasonal factor accounts for 20% of the weight, based on whether the current month is in the peak construction season; historical alarm frequency factor accounts for 30% of the weight, based on the number of alarms at the monitoring point in the past 30 days; construction registration factor accounts for 10% of the weight, based on whether there are valid construction registrations around the monitoring point. The scores of the four factors are added together and normalized to the range of 0-100.

[0009] In a preferred embodiment, in step S5, the key monitoring area includes a primary region of interest and a secondary region of interest. The primary region of interest is the corresponding area of ​​the cable channel in the image coordinate system, and the secondary region of interest is the area formed by extending the primary region of interest outward by a preset range. The detection frequency of the primary region of interest is twice that of the secondary region of interest.

[0010] In a preferred embodiment, in step S7, the formula for calculating the actual distance D is: In the formula, GD is the installation height of the camera, JJ is the focal length of the camera, MG is the height of the target detection box in the image, SK is the actual width of the reference target, TK is the width of the target detection box in the image, and BC is the installation angle compensation coefficient. When the actual distance D is less than the preset warning distance, the target is determined to have entered the warning area.

[0011] In a preferred embodiment, in step S8, the Kalman filter algorithm is used to smoothly track the position of the target in multiple consecutive frames of images. The motion velocity vector and motion direction are calculated based on the position change of the target in N consecutive frames. When the angle between the target's motion direction and the center line direction of the cable channel is less than a preset angle threshold, it is determined that the motion trend points to the cable channel.

[0012] In a preferred embodiment, in step S9, after receiving the alarm information, the background management platform queries whether there are valid construction registration records around the monitoring point. If there are valid registration records and the construction time and scope match the alarm information, the alarm is marked as registered construction and no alarm notification is pushed to the operation and maintenance personnel. If there are no matching registration records, the alarm is marked as suspected external damage and an alarm notification is immediately pushed to the operation and maintenance personnel.

[0013] An AI-powered intelligent monitoring system for preventing external damage to cable channels, based on cameras and solar energy, includes: Solar panels are used to convert light energy into electrical energy to provide power for operation. The energy storage unit is electrically connected to the solar panel and is used to store electrical energy and provide power. The power management unit is used to monitor the remaining power of the energy storage battery and the power generation of the solar photovoltaic panel in real time, calculate the sustainable working time CX, and dynamically adjust the working mode according to the sustainable working time CX. The acquisition unit is used to acquire video images of the cable channel and surrounding area. Its acquisition frame rate is adaptively adjusted according to the risk level value Rt of the current monitored area. The delineation unit is used to dynamically delineate key monitoring areas in the acquired video images based on the cable channel's direction, width, and the camera's current installation orientation. The identification unit has a built-in lightweight target detection model, which is used to perform real-time analysis of video images in key monitoring areas and identify externally damaged targets in the images; The distance unit is used to calculate the actual distance D between the target and the cable channel when an external damage risk target is identified, and to determine whether the target has entered the warning area. The determination unit is used to continuously track targets entering the warning area, calculate the target's velocity vector and direction of motion, and determine whether the target's direction of motion points to the center line of the cable channel. The alarm unit is used to determine the risk of external damage when a target enters the warning area and its movement trend points towards the cable channel, and to trigger an audible and visual alarm. The back-end management platform is used to receive alarm information, store and analyze data, and issue configuration parameters.

[0014] The technical effects and advantages of this invention are as follows: 1. This invention incorporates the current remaining power of the energy storage battery and the real-time power generation of the solar photovoltaic panel into the sustainable working time CX, and proactively estimates the sustainable working time. Based on the estimation results, it automatically switches between normal working mode, low power consumption mode and sleep mode. When there is sufficient sunlight, it makes full use of solar energy to operate in full function, and when there is insufficient sunlight, it intelligently saves power to extend the battery life, solves the problem of power supply difficulties for monitoring equipment in remote areas, and realizes all-weather active monitoring capability. 2. This invention deploys a lightweight target detection model locally, completing target detection inference locally, reducing recognition latency to milliseconds, eliminating reliance on network transmission, and using an adaptive region of interest strategy to detect only key areas, reducing unnecessary computation. By calculating the actual distance D between the target and the cable channel, it achieves accurate identification and positioning of external damage risk targets such as excavators, cranes, and pile drivers, reducing false alarm and false negative rates. 3. This invention comprehensively assesses the risk level based on four dimensions: time period, season, historical alarm frequency, and construction record. It maps the risk level to the specific current image acquisition frame rate (PL), enabling high-frequency acquisition in high-risk areas and low-frequency acquisition in low-risk areas. While ensuring monitoring accuracy during high-risk periods, it reduces system energy consumption during low-risk periods. It also works in conjunction with dynamic energy efficiency management strategies to extend battery life. 4. This invention dynamically adjusts the monitoring range according to the cable channel direction through an adaptive region of interest strategy, and marks legally registered construction through a construction registration verification mechanism, thus avoiding invalid alarms and improving the accuracy and practicality of alarms. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the intelligent monitoring method of the present invention.

[0016] Figure 2 This is a schematic diagram of the intelligent monitoring system of the present invention. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The AI ​​intelligent monitoring method and system for preventing external damage to cable channels based on cameras and solar energy involved in the present invention are not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Reference Figure 1 This invention provides an AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, comprising the following steps: Step S1: Deploy monitoring devices above or to the side of the cable channel, and configure monitoring areas, warning areas, and alarm parameters; Step S2: The solar photovoltaic panel converts light energy into electrical energy and stores it in the energy storage battery. The remaining power of the energy storage battery and the power generation of the solar photovoltaic panel are monitored in real time. Step S3: Calculate the sustainable working time CX and adjust the mode accordingly; Step S4: Calculate the current image acquisition frame rate PL; Step S5: Based on the cable channel's direction and width, as well as the camera's current installation posture, dynamically delineate key monitoring areas in the acquired video images, and only perform target detection and recognition processing on images within the key monitoring areas; Step S6: Perform real-time analysis on video images within the key monitoring area, and call the lightweight target detection model to identify external damage risk targets in the images. The external damage risk targets include excavators, cranes, and pile drivers. Step S7: When an external damage risk target is identified, calculate the actual distance D between the target and the cable channel; Step S8: Continuously track the target entering the warning area, calculate the target's velocity vector and direction of movement, and determine whether the target's direction of movement points towards the center line of the cable channel. When the target enters the warning area and its movement trend points towards the cable channel, it is determined that there is a risk of external damage. Step S9: After determining that there is a risk of external damage, issue an on-site audible and visual alarm and upload the alarm information to the back-end management platform.

[0019] The lightweight object detection model is obtained by compressing a large model in the cloud through knowledge distillation technology and performs object detection inference locally.

[0020] In this embodiment, the application provides all-weather intelligent monitoring of targets at risk of external damage around cable channels. Solar power combined with calculation of sustainable working time solves the power supply problem in the field where there is no mains power. The delineation of key areas reduces the amount of invalid calculation at the edge. AI local recognition avoids the network latency problem of cloud recognition. The dual verification of distance estimation and movement trend ensures the accuracy of alarms, thus forming a complete intelligent monitoring technology solution for preventing external damage to cable channels.

[0021] Reference Figure 1 In step S3, the formula for calculating the sustainable working time CX is as follows: In the formula, SY represents the current remaining capacity of the energy storage battery, GP represents the current measured power generation of the solar photovoltaic panel, and XL represents the charging conversion efficiency. YL is the preset prediction time window, GH is the reserved safe power, and GH is the system power consumption in the current working mode. In step S3, the calculated sustainable working time CX is compared with the preset first threshold Y1 and second threshold Y2, and the first threshold Y1 > the second threshold Y2. When the sustainable working time CX > the first threshold Y1, the normal working mode is maintained and the camera continuously collects images. When the first threshold Y1 ≥ the sustainable working time CX > the second threshold Y2, the low power mode is switched and the camera changes to interval acquisition. When the second threshold Y2 ≥ the sustainable working time, the sleep mode is switched and the camera stops working, only retaining the wireless communication function.

[0022] In this embodiment, the sustainable operating time CX incorporates two dynamic variables: the current remaining power of the energy storage battery SY and the real-time power generation of the solar photovoltaic panel GP. This forward-looking estimation of the sustainable operating time enables energy efficiency scheduling decisions to have both current state awareness and future trend prediction capabilities, effectively addressing complex scenarios such as sudden weather changes and continuous rain. It can operate continuously under unstable power supply conditions, and the operating state is finely divided into three levels: normal operating mode, low power consumption mode, and sleep mode, realizing gradient management of energy consumption. Compared with the existing technology that only has a two-state switching of working / sleep, the three-level gradient switching avoids the impact of frequent start-stop on the equipment. At the same time, when the sustainable operating time CX is in the middle range, the monitoring continuity is maintained by reducing the frame rate instead of directly shutting down, achieving the optimal balance between power supply endurance and monitoring effect.

[0023] Reference Figure 1 In step S4, the formula for calculating the current image acquisition frame rate PL is as follows: In the formula The minimum capture frame rate, The highest capture frame rate, This represents the current risk level value for the monitored area. For the highest risk level, the camera collects video images of the cable channel and surrounding area according to the calculated current image acquisition frame rate PL. In step S4, the risk level is determined as follows: time period factor accounts for 40% of the weight, based on whether the current period is within a preset high-incidence construction period; seasonal factor accounts for 20% of the weight, based on whether the current month is within the peak construction season; historical alarm frequency factor accounts for 30% of the weight, based on the number of alarms at the monitoring point in the past 30 days; construction registration factor accounts for 10% of the weight, based on whether there are valid construction registrations around the monitoring point. The scores of the four factors are added together and normalized to the range of 0-100.

[0024] In this embodiment of the application, the frame rate and risk level value are collected. Dynamic correlation enables adaptive scheduling of high-risk, high-frequency data collection and low-risk, low-frequency data collection. The formula adopts a linear mapping method, which has clear physical meaning and low computational load. It reduces overall energy consumption while ensuring monitoring accuracy during high-risk periods. Furthermore, this application comprehensively assesses the risk level from four dimensions: time period, season, historical alarm frequency, and construction record. The weighted integration of these four factors can accurately perceive the differentiated risk levels of different monitoring points and time periods, providing a reliable data foundation for adaptive frame rate data collection.

[0025] Reference Figure 1In step S5, the key monitoring area includes a primary region of interest and a secondary region of interest. The primary region of interest is the corresponding area of ​​the cable channel in the image coordinate system, and the secondary region of interest is the area formed by extending the primary region of interest outward by a preset range. The detection frequency of the primary region of interest is twice that of the secondary region of interest.

[0026] In this embodiment, a primary region of interest is dynamically generated based on the actual route of the cable channel, and then expanded outward to form a secondary region of interest. The two regions are layered with a difference of twice the detection frequency. Computational resources are concentrated in the core area of ​​the cable channel, which not only ensures high-frequency monitoring of key areas, but also provides an early warning buffer for target approach through the secondary region of interest, thereby reducing the amount of invalid calculation while ensuring the monitoring effect.

[0027] Reference Figure 1 In step S7, the formula for calculating the actual distance D is as follows: In the formula, GD is the installation height of the camera, JJ is the focal length of the camera, MG is the height of the target detection box in the image, SK is the actual width of the reference target, TK is the width of the target detection box in the image, and BC is the installation angle compensation coefficient. When the actual distance D is less than the preset warning distance, the target is determined to have entered the warning area.

[0028] In this embodiment, the application utilizes the principle of monocular vision ranging, combined with the camera installation height GD, focal length JJ, target size in the image (MG, TK), and reference target actual width SK, to accurately calculate the actual distance D between the target and the cable channel, thereby elevating the alert determination from two-dimensional image space to three-dimensional real space and significantly improving the accuracy of alert determination.

[0029] Reference Figure 1 In step S8, the Kalman filter algorithm is used to smoothly track the position of the target in multiple consecutive frames of images. The motion velocity vector and motion direction are calculated based on the position change of the target in N consecutive frames. When the angle between the target's motion direction and the center line direction of the cable channel is less than a preset angle threshold, it is determined that the motion trend points to the cable channel.

[0030] In this embodiment, the Kalman filter algorithm is used to smoothly track the position of the target in multiple consecutive frames, effectively filtering out detection noise. The motion trend is determined by calculating the angle between the motion direction and the center line of the cable channel. An alarm is triggered only when the angle is less than a preset angle threshold. This can accurately identify dangerous targets approaching the cable channel, eliminate interference factors such as passing vehicles and pedestrians, and reduce the false alarm rate.

[0031] Reference Figure 1In step S9, after receiving the alarm information, the backend management platform queries whether there are valid construction filing records around the monitoring point. If there are valid filings and the construction time and scope match the alarm information, the alarm is marked as filed construction and no alarm notification is pushed to the operation and maintenance personnel. If there are no matching filing records, the alarm is marked as suspected external damage and an alarm notification is immediately pushed to the operation and maintenance personnel.

[0032] In this embodiment, the construction registration records are queried in real time through the back-end management platform. Alarms that have been registered and whose construction time and scope match are marked and not pushed to the system. Alarms that do not match the registration are pushed to the system immediately. This achieves the goal of not alarming for legally registered construction and immediately alarming for suspected external damage that has not been registered. This improves the accuracy of alarm push and allows maintenance personnel to focus their attention on dangerous situations that really need to be dealt with, thereby improving maintenance efficiency.

[0033] Reference Figure 2 An AI-powered intelligent monitoring system for preventing external damage to cable channels, based on cameras and solar energy, includes: Solar panels are used to convert light energy into electrical energy to provide power for operation. The energy storage unit is electrically connected to the solar panel and is used to store electrical energy and provide power. The power management unit is used to monitor the remaining power of the energy storage battery and the power generation of the solar photovoltaic panel in real time, calculate the sustainable working time CX, and dynamically adjust the working mode according to the sustainable working time CX. The acquisition unit is used to acquire video images of the cable channel and surrounding area. Its acquisition frame rate is adaptively adjusted according to the risk level value Rt of the current monitored area. The delineation unit is used to dynamically delineate key monitoring areas in the acquired video images based on the cable channel's direction, width, and the camera's current installation orientation. The identification unit has a built-in lightweight target detection model, which is used to perform real-time analysis of video images in key monitoring areas and identify externally damaged targets in the images; The distance unit is used to calculate the actual distance D between the target and the cable channel when an external damage risk target is identified, and to determine whether the target has entered the warning area. The determination unit is used to continuously track targets entering the warning area, calculate the target's velocity vector and direction of motion, and determine whether the target's direction of motion points to the center line of the cable channel. The alarm unit is used to determine the risk of external damage when a target enters the warning area and its movement trend points towards the cable channel, and to trigger an audible and visual alarm. The back-end management platform is used to receive alarm information, store and analyze data, and issue configuration parameters.

[0034] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. The units and algorithm steps of the various examples described in the embodiments can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0035] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0036] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0037] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, characterized in that: Includes the following steps: Step S1: Deploy monitoring devices above or to the side of the cable channel, and configure monitoring areas, warning areas, and alarm parameters; Step S2: The solar photovoltaic panel converts light energy into electrical energy and stores it in the energy storage battery. The remaining power of the energy storage battery and the power generation of the solar photovoltaic panel are monitored in real time. Step S3: Calculate the sustainable working time CX and adjust the mode accordingly; Step S4: Calculate the current image acquisition frame rate PL; Step S5: Based on the cable channel's direction and width, as well as the camera's current installation posture, dynamically delineate key monitoring areas in the acquired video images, and only perform target detection and recognition processing on images within the key monitoring areas; Step S6: Perform real-time analysis on video images within the key monitoring area, and call the lightweight target detection model to identify external damage risk targets in the images. The external damage risk targets include excavators, cranes, and pile drivers. Step S7: When an external damage risk target is identified, calculate the actual distance D between the target and the cable channel; Step S8: Continuously track the target entering the warning area, calculate the target's velocity vector and direction of movement, and determine whether the target's direction of movement points towards the center line of the cable channel. When the target enters the warning area and its movement trend points towards the cable channel, it is determined that there is a risk of external damage. Step S9: After determining that there is a risk of external damage, issue an on-site audible and visual alarm and upload the alarm information to the back-end management platform.

2. The AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, as described in claim 1, is characterized in that: In step S3, the formula for calculating the sustainable working time CX is as follows: In the formula, SY represents the current remaining capacity of the energy storage battery, GP represents the current measured power generation of the solar photovoltaic panel, and XL represents the charging conversion efficiency. YL represents the preset prediction time window, YL represents the reserved safe power, and GH represents the system power consumption in the current operating mode.

3. The AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, as described in claim 1, is characterized in that: In step S3, the calculated continuous working time CX is compared with the preset first threshold Y1 and second threshold Y2, and the first threshold Y1 > the second threshold Y2. When the continuous working time CX > the first threshold Y1, the normal working mode is maintained and the camera continuously collects images. When the first threshold Y1 ≥ the continuous working time CX > the second threshold Y2, the low power mode is switched and the camera changes to interval acquisition. When the second threshold Y2 ≥ the continuous working time, the sleep mode is switched and the camera stops working, only retaining the wireless communication function.

4. The AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, as described in claim 1, is characterized in that: In step S4, the formula for calculating the current image acquisition frame rate PL is: In the formula The minimum capture frame rate, The highest capture frame rate, This represents the current risk level value for the monitored area. At the highest risk level, the camera captures video images of the cable channel and surrounding area according to the calculated current image acquisition frame rate PL.

5. The AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, as described in claim 1, is characterized in that: In step S4, the risk level value is determined as follows: time period factor accounts for 40% of the weight, based on whether the current period is within a preset high-incidence construction period; seasonal factor accounts for 20% of the weight, based on whether the current month is within the peak construction season; historical alarm frequency factor accounts for 30% of the weight, based on the number of alarms at the monitoring point in the past 30 days; construction registration factor accounts for 10% of the weight, based on whether there are valid construction registrations around the monitoring point. The scores of the four factors are added together and normalized to the range of 0-100.

6. The AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, as described in claim 1, is characterized in that: In step S5, the key monitoring area includes a primary region of interest and a secondary region of interest. The primary region of interest is the corresponding area of ​​the cable channel in the image coordinate system, and the secondary region of interest is the area formed by extending the primary region of interest outward by a preset range. The detection frequency of the primary region of interest is twice that of the secondary region of interest.

7. The AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, as described in claim 1, is characterized in that: In step S7, the formula for calculating the actual distance D is: In the formula, GD is the installation height of the camera, JJ is the focal length of the camera, MG is the height of the target detection box in the image, SK is the actual width of the reference target, TK is the width of the target detection box in the image, and BC is the installation angle compensation coefficient. When the actual distance D is less than the preset warning distance, the target is determined to have entered the warning area.

8. The AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, as described in claim 1, is characterized in that: In step S8, the Kalman filter algorithm is used to smoothly track the position of the target in multiple consecutive frames of images. The motion velocity vector and motion direction are calculated based on the position change of the target in N consecutive frames. When the angle between the target's motion direction and the center line of the cable channel is less than a preset angle threshold, it is determined that the motion trend points to the cable channel.

9. The AI-powered intelligent monitoring method for preventing external damage to cable channels based on cameras and solar energy, as described in claim 1, is characterized in that: In step S9, after receiving the alarm information, the backend management platform queries whether there are valid construction registration records around the monitoring point. If there are valid registration records and the construction time and scope match the alarm information, the alarm is marked as registered construction and no alarm notification is pushed to the operation and maintenance personnel. If there are no matching registration records, the alarm is marked as suspected external damage and an alarm notification is immediately pushed to the operation and maintenance personnel.

10. An AI-powered intelligent monitoring system for preventing external damage to cable channels based on cameras and solar energy, characterized in that: include: Solar panels are used to convert light energy into electrical energy to provide power for operation. The energy storage unit is electrically connected to the solar panel and is used to store electrical energy and provide power. The power management unit is used to monitor the remaining power of the energy storage battery and the power generation of the solar photovoltaic panel in real time, calculate the sustainable working time CX, and dynamically adjust the working mode according to the sustainable working time CX. The acquisition unit is used to acquire video images of the cable channel and surrounding area. Its acquisition frame rate is adaptively adjusted according to the risk level value Rt of the current monitored area. The delineation unit is used to dynamically delineate key monitoring areas in the acquired video images based on the cable channel's direction, width, and the camera's current installation orientation. The identification unit has a built-in lightweight target detection model, which is used to perform real-time analysis of video images in key monitoring areas and identify externally damaged targets in the images; The distance unit is used to calculate the actual distance D between the target and the cable channel when an external damage risk target is identified, and to determine whether the target has entered the warning area. The determination unit is used to continuously track targets entering the warning area, calculate the target's velocity vector and direction of motion, and determine whether the target's direction of motion points to the center line of the cable channel. The alarm unit is used to determine the risk of external damage when a target enters the warning area and its movement trend points towards the cable channel, and to trigger an audible and visual alarm. The back-end management platform is used to receive alarm information, store and analyze data, and issue configuration parameters.

Images