Low-power-consumption dam termite monitoring system and method based on intelligent image recognition

Abstract

The invention relates to the technical field of dam safety monitoring, in particular to a low-power-consumption dam termite monitoring system and method based on intelligent image recognition. According to the technical scheme, the low-power-consumption dam termite monitoring system and method based on intelligent image recognition comprises a data analysis module; compared with a traditional termite monitoring system and a manual patrol mode which are applied for a long time, the termite monitoring system has the obvious defects of high system power consumption, difficulty in stable operation in remote areas such as dams and the like for a long time, poor environmental adaptability, great reduction of the recognition effect especially when light is insufficient or weather conditions are poor, and the like. The termite monitoring system can continuously collect high-quality images under the condition that power consumption is limited, and is particularly suitable for long-time unattended environment monitoring of dams and the like. The system is ensured to realize precise monitoring and real-time early warning of termite activities under the condition of low energy consumption, and meanwhile, a termite image recognition model of deep learning and small target detection is constructed, so that accurate and efficient extraction of termite information is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of dam safety monitoring, and in particular to a low-power dam termite monitoring system and method based on intelligent image recognition. Background Art

[0002] As an important water conservancy engineering facility, the safety and stability of dams are directly related to the safety of people’s lives and property and the sustainable development of the social economy.

[0003] The existing termite monitoring system, although the manual inspection method has been used for a long time, has obvious shortcomings. The system has high power consumption and is difficult to operate stably for a long time in remote areas such as dams. It has poor environmental adaptability, especially in insufficient light or bad weather conditions. The recognition effect is greatly reduced. At the same time, the existing monitoring data cannot meet the requirements of accurate early warning, and it is difficult to accurately identify the characteristics of termite activity from a large amount of data. At the same time, the system cannot intelligently distinguish between equipment failures and termite damage, and lacks a scientific early warning system and emergency response mechanism, resulting in the inability to predict termite risks in advance and take timely measures.

[0004] For the existing termite monitoring system, although the manual patrol method has been used for a long time, it has obvious shortcomings. The system power consumption is high, and it is difficult to operate stably for a long time in remote areas such as dams, and the environmental adaptability is poor. The termite monitoring system can continuously collect high-quality images under limited power consumption. It is particularly suitable for long-term unattended environmental monitoring such as dams, ensuring that the system can achieve accurate monitoring and real-time early warning of termite activities under low energy consumption conditions. At the same time, a termite image recognition model based on deep learning and small target detection is constructed to achieve accurate and efficient extraction of termite information. Summary of the invention

[0005] In order to overcome the existing termite monitoring system, although the manual patrol method has been used for a long time, it has obvious shortcomings. The system has high power consumption and is difficult to operate stably for a long time in remote areas such as dams. It has poor environmental adaptability, especially in insufficient light or bad weather conditions. The recognition effect is greatly reduced. At the same time, the existing monitoring data cannot meet the requirements of accurate early warning, and it is difficult to accurately identify the characteristics of termite activity from a large amount of data. At the same time, the system cannot intelligently distinguish between equipment failures and termite damage, and lacks a scientific early warning system and emergency response mechanism, resulting in the inability to predict termite risks in advance and take timely measures.

[0006] The technical solution of the present invention is: a low-power dam termite monitoring system based on intelligent image recognition, comprising: Data acquisition module: used to collect image data of termite images; Data receiving module: used to receive the termite image data collected by the data collection module; Data analysis module: used for real-time analysis of termite image data received by the data receiving module; Real-time monitoring and early warning module: used to detect data in real time to confirm termite activities; System adaptation and remote maintenance module: used for remote upgrade of monitoring devices, parameter setting and self-diagnosis of equipment failures.

[0007] Preferably, the data acquisition module includes multiple groups of low-power cameras, an embedded core processor and a power module. The data acquisition module is used to input data from the termite monitoring device to ensure that the monitoring data can be collected and transmitted in real time. It also supports a variety of heterogeneous sensor protocols and communication methods, can adapt to monitoring data from different sources, and realize real-time collection of termite image data.

[0008] Preferably, the data receiving module is used to access the data of the termite monitoring device. As the data entry of the entire monitoring system, the data receiving module can support a variety of heterogeneous sensor protocols and communication methods, ensuring that monitoring data from different sources and in different formats can be collected into the system in real time and accurately. At the same time, it has the functions of data format conversion, analysis and classified storage, and performs preliminary processing on the received monitoring data to provide reliable data support for subsequent intelligent analysis and early warning.

[0009] Preferably, the data analysis module is used to utilize big data and machine learning technology to perform real-time analysis of the received monitoring data. By comparing historical data with real-time monitoring data, the trend of termite activity is analyzed to confirm the presence and activity of termites. Once termite activity is detected, the system will immediately generate an alarm message and notify management personnel through multiple channels to ensure that the problem can be responded to and handled in a timely manner. According to environmental changes and monitoring results, the system will automatically optimize and adjust detection parameters to improve monitoring accuracy and energy-saving effects. At the same time, when the equipment fails or the monitoring results are abnormal, the system will perform fault diagnosis and generate a detailed fault report.

[0010] Preferably, the real-time monitoring and early warning module is used to monitor the data received from the termite monitoring device in real time and analyze the activities of termites. Once termite activities are detected, the module will automatically send early warning information to the management personnel to ensure that the problem can be responded to in time and improve the accuracy of the early warning.

[0011] Preferably, the system adaptation and remote maintenance module is used to automatically adjust the detection parameters and working mode according to the environmental conditions and monitoring results to improve the detection capability and energy-saving effect of the system. At the same time, it supports remote upgrades, parameter settings and self-diagnosis of equipment failures of the monitoring device. Once an equipment failure or monitoring abnormality occurs, the system will actively generate a report and report it to the management platform to ensure the stable operation of the monitoring device. Through adaptive adjustment and remote maintenance, the system can maintain an efficient and stable operating state, reduce monitoring errors caused by equipment failure or improper parameters, and provide strong guarantees for the long-term safety and stable operation of the dam.

[0012] Preferably, through low power consumption and efficient monitoring, the low power consumption camera can continuously collect high-quality images under limited power consumption conditions, which is particularly suitable for long-term unattended environmental monitoring such as dams, ensuring that the system can achieve accurate monitoring and real-time early warning of termite activities under low energy consumption conditions, greatly improving the sustainability of the system and the stability of monitoring. At the same time, real-time alarm and data management, the termite monitoring system integrates collection, intelligent analysis, directional early warning and other functions into a full-chain supervision system, realizing online management of termite prevention and control. When termite activity is detected, the system will issue a real-time alarm and notify the management personnel to ensure a rapid response to the problem.

[0013] A low-power dam termite monitoring method based on intelligent image recognition comprises the following steps: S11: Use low-power cameras to regularly collect images to accurately detect termites and provide real-time warnings; S12: Receive data through the data receiving module and upload the data from the edge to the cloud; S13: After receiving the real-time termite data, the data analysis module analyzes and calculates termite activities; S14: When termites are active, we combine big data and machine learning technology to confirm termite activity through historical data and real-time monitoring data, and automatically send warning information to relevant personnel and management platforms to ensure rapid response to problems; S15: The system adaptation and remote maintenance module automatically adjusts the detection parameters and working mode according to the environmental conditions and monitoring results to improve the system's detection capability and energy-saving effect.

[0014] Preferably, when the “cloud-edge” collaborative termite intelligent identification method is implemented through the edge end, the following steps are included: S21: The image of the dam area is collected regularly by a low-power camera, and the camera works continuously to ensure low energy consumption during long-term operation; S22: Read the image and use the cv2.imread() function to read the image in the specified path; S23: Further, the image is converted from the BGR color space to the HSV color space. The HSV color space is more suitable for color-based image segmentation because it decomposes color information into three components: hue, saturation, and brightness; S24: Set the color range and perform color filtering to define the approximate color range of termites in the HSV color space, and use the cv2.inRange() function to generate a binary image according to the defined color range, where the white area represents the possible termite area and the black area represents the non-termite area; S25: Use morphological closing operation to close small holes or small black spots inside foreground objects and connect neighboring objects, which helps to remove small noise points and make the termite area more continuous; S26: Gaussian blur is performed on the binary image after morphological processing to further reduce noise and uneven edges, and the Canny edge detection algorithm is applied to detect the edges in the image. This step is to extract clearer edge information from the blurred binary image to prepare for subsequent contour search; S27: Use cv2.findContours() function to find contours in the image.

[0015] Preferably, when the "cloud-edge" collaborative termite intelligent identification method is implemented through the cloud, the following steps are included: S31: The cloud receives images uploaded by the edge, stores and classifies the images, further detects the uploaded images through deep learning algorithms, and combines historical data and large-scale computing power to perform more sophisticated termite activity identification; S32: Cloud detection result processing, if termites are detected in the cloud, send warning information to the management personnel, and record relevant data; S33: There are two situations in which termites are not detected: termites have been detected on the edge, but not on the cloud. A manual review request is sent to ensure that the abnormal situation is reviewed. If termites are not detected on the edge and not on the cloud, the image is transferred to the storage area for subsequent analysis and evaluation.

[0016] Preferably, in the termite detection system, when the modules cooperate with each other, the following steps are included: S41: The system accesses the data of the termite monitoring device through the data receiving module, supports a variety of heterogeneous sensor protocols and communication methods, such as wireless communication, Internet of Things protocols, etc., to ensure that monitoring data from different sources can be collected in real time. The data receiving module converts the format of the collected monitoring data to ensure that the data can adapt to the standardized storage and analysis format of the system. The system performs a preliminary analysis of the received monitoring data to determine the validity and quality of the data, and then classifies and stores the data in a local database for subsequent analysis and processing; S42: The real-time monitoring and early warning module compares historical monitoring data with real-time collected data through big data analysis and machine learning algorithms to analyze termite activity trends. Once termite activity is detected, the system will immediately generate an alarm message and notify management personnel through channels such as phone calls, text messages, and mobile phone apps to ensure that the problem can be responded to and handled in a timely manner; S43: The system has an adaptive adjustment function. According to environmental changes and monitoring results, it automatically optimizes and adjusts detection parameters, such as acquisition frequency, sensitivity, etc., to improve monitoring accuracy and energy saving effects. When a device fails or the monitoring results are abnormal, the system will automatically detect and generate a detailed fault report, and then upload the report to the management platform through the remote maintenance module. After receiving the equipment fault report, the management platform can remotely diagnose the problem and perform maintenance.

[0017] Beneficial effects of the present invention: 1. Compared with the traditional termite monitoring system, although the manual inspection method has been used for a long time, it has obvious shortcomings. The system has high power consumption and is difficult to operate stably in remote areas such as dams for a long time. The environmental adaptability is poor, especially when there is insufficient light or bad weather conditions, the recognition effect is greatly reduced. At the same time, the existing monitoring data cannot meet the requirements of accurate early warning, and it is difficult to accurately identify the characteristics of termite activities from a large amount of data. At the same time, the system cannot intelligently distinguish between equipment failure and termite damage, lacks a scientific early warning system and emergency response mechanism, resulting in the inability to predict termite risks in advance and take timely measures. The termite monitoring system can continuously collect high-quality images under limited power consumption, which is particularly suitable for long-term unattended environmental monitoring such as dams, ensuring that the system can achieve accurate monitoring and real-time early warning of termite activities under low energy consumption conditions. At the same time, a termite image recognition model based on deep learning and small target detection is constructed to achieve accurate and efficient extraction of termite information; 2. Through low power consumption and efficient monitoring, low-power cameras can continuously collect high-quality images under limited power consumption conditions. They are particularly suitable for long-term unattended environmental monitoring such as dams, ensuring that the system can accurately monitor and warn termites in real time under low energy conditions, greatly improving the sustainability of the system and the stability of monitoring. At the same time, real-time alarms and data management, the termite monitoring system integrates data collection, intelligent analysis, directional warning and other functions into a full-chain supervision system to achieve online management of termite prevention and control. When termite activity is detected, the system will issue a real-time alarm and notify management personnel to ensure a rapid response to the problem. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 Shown is a schematic diagram of a "cloud-edge" collaborative termite intelligent identification method of a low-power dam termite monitoring system based on intelligent image recognition of the present invention; Figure 2What is shown is a flow chart of a low-power dam termite monitoring system and method based on intelligent image recognition of the present invention; Figure 3 Shown is a schematic diagram of the workflow of a low-power dam termite monitoring method based on intelligent image recognition of the present invention. DETAILED DESCRIPTION

[0019] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0020] See also Figure 1-3 The present invention provides an embodiment: a low-power dam termite monitoring system based on intelligent image recognition, comprising: Data acquisition module: used to collect image data of termite images; Data receiving module: used to receive the termite image data collected by the data collection module; Data analysis module: used for real-time analysis of termite image data received by the data receiving module; Real-time monitoring and early warning module: used to detect data in real time to confirm termite activities; System adaptation and remote maintenance module: used for remote upgrade of monitoring devices, parameter setting and self-diagnosis of equipment failures.

[0021] Preferably, the data acquisition module includes multiple groups of low-power cameras, an embedded core processor and a power module. The data acquisition module is used to input data from the termite monitoring device to ensure that the monitoring data can be collected and transmitted in real time. It also supports a variety of heterogeneous sensor protocols and communication methods, can adapt to monitoring data from different sources, and realize real-time collection of termite image data.

[0022] Preferably, the data receiving module is used to access the data of the termite monitoring device. As the data entry of the entire monitoring system, the data receiving module can support a variety of heterogeneous sensor protocols and communication methods, ensuring that monitoring data from different sources and in different formats can be collected into the system in real time and accurately. At the same time, it has the functions of data format conversion, analysis and classified storage, and performs preliminary processing on the received monitoring data to provide reliable data support for subsequent intelligent analysis and early warning.

[0023] Preferably, the data analysis module is used to utilize big data and machine learning technology to perform real-time analysis of the received monitoring data. By comparing historical data with real-time monitoring data, the trend of termite activity is analyzed to confirm the presence and activity of termites. Once termite activity is detected, the system will immediately generate an alarm message and notify management personnel through multiple channels to ensure that the problem can be responded to and handled in a timely manner. According to environmental changes and monitoring results, the system will automatically optimize and adjust detection parameters to improve monitoring accuracy and energy-saving effects. At the same time, when the equipment fails or the monitoring results are abnormal, the system will perform fault diagnosis and generate a detailed fault report.

[0024] Preferably, the real-time monitoring and early warning module is used to monitor the data received from the termite monitoring device in real time and analyze the activities of termites. Once termite activities are detected, the module will automatically send early warning information to the management personnel to ensure that the problem can be responded to in time and improve the accuracy of the early warning.

[0025] Preferably, the system adaptation and remote maintenance module is used to automatically adjust the detection parameters and working mode according to the environmental conditions and monitoring results to improve the detection capability and energy-saving effect of the system. At the same time, it supports remote upgrades, parameter settings and self-diagnosis of equipment failures of the monitoring device. Once an equipment failure or monitoring abnormality occurs, the system will actively generate a report and report it to the management platform to ensure the stable operation of the monitoring device. Through adaptive adjustment and remote maintenance, the system can maintain an efficient and stable operating state, reduce monitoring errors caused by equipment failure or improper parameters, and provide strong guarantees for the long-term safety and stable operation of the dam.

[0026] Optimally, through low power consumption and efficient monitoring, low-power cameras can continuously collect high-quality images under limited power consumption, which is particularly suitable for long-term unattended environmental monitoring such as dams, ensuring that the system can accurately monitor termite activities and provide real-time warnings under low energy consumption conditions, greatly improving the sustainability of the system and the stability of monitoring. At the same time, real-time alarms and data management, the termite monitoring system integrates collection, intelligent analysis, directional warning and other functions into a full-chain supervision system to achieve online management of termite prevention and control. When termite activity is detected, the system will issue a real-time alarm and notify the management personnel to ensure a quick response to the problem.

[0027] A low-power dam termite monitoring method based on intelligent image recognition comprises the following steps: S11: Use low-power cameras to regularly collect images to accurately detect termites and provide real-time warnings; S12: Receive data through the data receiving module and upload the data from the edge to the cloud; S13: After receiving the real-time termite data, the data analysis module analyzes and calculates termite activities; S14: When termites are active, we combine big data and machine learning technology to confirm termite activity through historical data and real-time monitoring data, and automatically send warning information to relevant personnel and management platforms to ensure rapid response to problems; S15: The system adaptation and remote maintenance module automatically adjusts the detection parameters and working mode according to the environmental conditions and monitoring results to improve the system's detection capability and energy-saving effect.

[0028] Preferably, when the “cloud-edge” collaborative termite intelligent identification method is implemented through the edge end, the following steps are included: S21: The image of the dam area is collected regularly by a low-power camera, and the camera works continuously to ensure low energy consumption during long-term operation; S22: Read the image and use the cv2.imread() function to read the image in the specified path; S23: Further, the image is converted from the BGR color space to the HSV color space. The HSV color space is more suitable for color-based image segmentation because it decomposes color information into three components: hue, saturation, and brightness; S24: Set the color range and perform color filtering to define the approximate color range of termites in the HSV color space, and use the cv2.inRange() function to generate a binary image according to the defined color range, where the white area represents the possible termite area and the black area represents the non-termite area; S25: Use morphological closing operation to close small holes or small black spots inside foreground objects and connect neighboring objects. This helps remove small noise points and make the termite area more continuous; S26: Gaussian blur is performed on the binary image after morphological processing to further reduce noise and uneven edges, and the Canny edge detection algorithm is applied to detect the edges in the image. This step is to extract clearer edge information from the blurred binary image to prepare for subsequent contour search; S27: Use cv2.findContours() function to find contours in the image.

[0029] Preferably, when the "cloud-edge" collaborative termite intelligent identification method is implemented through the cloud, the following steps are included: S31: The cloud receives images uploaded by the edge, stores and classifies the images, further detects the uploaded images through deep learning algorithms, and combines historical data and large-scale computing power to perform more sophisticated termite activity identification; S32: Cloud detection result processing, if termites are detected in the cloud, send warning information to management personnel, and record relevant data; S33: There are two situations in which termites are not detected: termites have been detected on the edge, but not on the cloud. A manual review request is sent to ensure that the abnormal situation is reviewed. If termites are not detected on the edge and not on the cloud, the image is transferred to the storage area for subsequent analysis and evaluation.

[0030] Preferably, in the termite detection system, when the modules cooperate with each other, the following steps are included: S41: The system accesses the data of the termite monitoring device through the data receiving module, supports a variety of heterogeneous sensor protocols and communication methods, such as wireless communication, Internet of Things protocols, etc., to ensure that monitoring data from different sources can be collected in real time. The data receiving module converts the format of the collected monitoring data to ensure that the data can adapt to the standardized storage and analysis format of the system. The system performs a preliminary analysis of the received monitoring data to determine the validity and quality of the data, and then classifies and stores the data in a local database for subsequent analysis and processing; S42: The real-time monitoring and early warning module compares historical monitoring data with real-time collected data through big data analysis and machine learning algorithms to analyze termite activity trends. Once termite activity is detected, the system will immediately generate an alarm message and notify management personnel through channels such as phone calls, text messages, and mobile phone apps to ensure that the problem can be responded to and handled in a timely manner; S43: The system has an adaptive adjustment function. According to environmental changes and monitoring results, it automatically optimizes and adjusts detection parameters, such as acquisition frequency, sensitivity, etc., to improve monitoring accuracy and energy saving effects. When a device fails or the monitoring results are abnormal, the system will automatically detect and generate a detailed fault report, and then upload the report to the management platform through the remote maintenance module. After receiving the equipment fault report, the management platform can remotely diagnose the problem and perform maintenance.

[0031] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of those skilled in the art without departing from the spirit of the present invention.

Claims

1. A low-power dam termite monitoring system based on intelligent image recognition; characterized by: Included are: Data acquisition module: used to collect image data of termite images. The data acquisition module includes multiple sets of low-power cameras, an embedded core processor and a power module; Data receiving module: used to receive the termite image data collected by the data collection module; Data analysis module: used to perform real-time analysis on the termite image data received by the data receiving module; used to use big data and machine learning technology to perform real-time analysis on the received monitoring data, analyze the trend of termite activity by comparing historical data with real-time monitoring data, and confirm the existence and activity of termites. Once termite activity is detected, the system will immediately generate an alarm message and notify the management personnel through multiple channels to ensure that the problem can be responded to and handled in a timely manner; Real-time monitoring and early warning module: used to detect data in real time to confirm termite activities; System adaptation and remote maintenance module: used for remote upgrade of monitoring devices, parameter setting and self-diagnosis of equipment failures.

2. The low-power dam termite monitoring system based on intelligent image recognition according to claim 1 is characterized in that: The data acquisition module is used to input data from the termite monitoring device to ensure that the monitoring data can be collected and transmitted in real time. It also supports a variety of heterogeneous sensor protocols and communication methods, can adapt to monitoring data from different sources, and realize real-time collection of termite image data.

3. The low-power dam termite monitoring system based on intelligent image recognition according to claim 2 is characterized by: The data receiving module is used to access the data of the termite monitoring device. As the data entry of the entire monitoring system, the data receiving module can support a variety of heterogeneous sensor protocols and communication methods, ensuring that monitoring data from different sources and in different formats can be collected into the system in real time and accurately. At the same time, it has the functions of data format conversion, analysis and classified storage, and performs preliminary processing on the received monitoring data to provide reliable data support for subsequent intelligent analysis and early warning.

4. The low-power dam termite monitoring system based on intelligent image recognition according to claim 3 is characterized by: The data analysis module will automatically optimize and adjust the detection parameters according to environmental changes and monitoring results to improve monitoring accuracy and energy saving effects. At the same time, when a device fails or the monitoring results are abnormal, the system will perform fault diagnosis and generate a detailed fault report.

5. The low-power dam termite monitoring system based on intelligent image recognition according to claim 4 is characterized in that: The real-time monitoring and early warning module is used to monitor the data received from the termite monitoring device in real time and analyze the activities of termites. Once termite activities are detected, the module will automatically send early warning information to the management personnel to ensure that the problem can be responded to in a timely manner and improve the accuracy of the early warning.

6. The low-power dam termite monitoring system based on intelligent image recognition according to claim 5 is characterized by: The system adaptation and remote maintenance module is used to automatically adjust the detection parameters and working mode according to the environmental conditions and monitoring results to improve the system's detection capabilities and energy-saving effects. It also supports remote upgrades, parameter settings and self-diagnosis of equipment failures of the monitoring device. Once an equipment failure or monitoring anomaly occurs, the system will actively generate a report and report it to the management platform to ensure the stable operation of the monitoring device. Through adaptive adjustment and remote maintenance, the system can maintain an efficient and stable operating state, reduce monitoring errors caused by equipment failure or improper parameters, and provide strong guarantees for the long-term safety and stable operation of the dam.

7. A low-power dam termite monitoring method based on intelligent image recognition, characterized in that A low-power dam termite monitoring system based on intelligent image recognition according to claim 6 comprises the following steps: S11: Use low-power cameras to regularly collect images to accurately detect termites and provide real-time warnings; S12: Receive data through the data receiving module and upload the data from the edge to the cloud; S13: After receiving the real-time termite data, the data analysis module analyzes and calculates termite activities; S14: When termites are active, we combine big data and machine learning technology to confirm termite activity through historical data and real-time monitoring data, and automatically send warning information to relevant personnel and management platforms to ensure rapid response to problems; S15: The system adaptation and remote maintenance module automatically adjusts the detection parameters and working mode according to the environmental conditions and monitoring results to improve the system's detection capability and energy-saving effect.

8. The low-power dam termite monitoring method based on intelligent image recognition according to claim 7 is characterized by: When the "cloud-edge" collaborative termite intelligent identification method is implemented through the edge, the following steps are included: S21: The image of the dam area is collected regularly by a low-power camera, and the camera works continuously to ensure low energy consumption during long-term operation; S22: Read the image and use the cv2.imread() function to read the image in the specified path; S23: Further, the image is converted from the BGR color space to the HSV color space. The HSV color space is more suitable for color-based image segmentation because it decomposes color information into three components: hue, saturation, and brightness; S24: Set the color range and perform color filtering to define the approximate color range of termites in the HSV color space, and use the cv2.inRange() function to generate a binary image according to the defined color range, where the white area represents the possible termite area and the black area represents the non-termite area; S25: Use morphological closing operation to close small holes or small black spots inside the foreground object and connect neighboring objects, which helps to remove small noise points and make the termite area more continuous; S26: Gaussian blur is performed on the binary image after morphological processing to further reduce noise and uneven edges, and the Canny edge detection algorithm is applied to detect the edges in the image. This step is to extract clearer edge information from the blurred binary image to prepare for subsequent contour search; S27: Use cv2.findContours() function to find contours in the image.

9. The low-power dam termite monitoring method based on intelligent image recognition according to claim 7, characterized in that: When the "cloud-edge" collaborative termite intelligent identification method is implemented through the cloud, the following steps are included: S31: The cloud receives images uploaded by the edge, stores and classifies the images, further detects the uploaded images through deep learning algorithms, and combines historical data and large-scale computing power to perform more sophisticated termite activity identification; S32: Cloud detection result processing, if termites are detected in the cloud, send warning information to management personnel, and record relevant data; S33: There are two situations in which termites are not detected: termites have been detected on the edge, but not on the cloud. A manual review request is sent to ensure that the abnormal situation is reviewed. If termites are not detected on the edge and not on the cloud, the image is transferred to the storage area for subsequent analysis and evaluation.

10. The low-power dam termite monitoring method based on intelligent image recognition according to claim 7, characterized in that: In the termite detection system, when the modules cooperate with each other, the following steps are included: S41: The system accesses the data of the termite monitoring device through the data receiving module, supports a variety of heterogeneous sensor protocols and communication methods, such as wireless communication, Internet of Things protocols, etc., to ensure that monitoring data from different sources can be collected in real time. The data receiving module converts the format of the collected monitoring data to ensure that the data can adapt to the standardized storage and analysis format of the system. The system performs a preliminary analysis of the received monitoring data to determine the validity and quality of the data, and then classifies and stores the data in a local database for subsequent analysis and processing; S42: The real-time monitoring and early warning module compares historical monitoring data with real-time collected data through big data analysis and machine learning algorithms to analyze termite activity trends. Once termite activity is detected, the system will immediately generate an alarm message and notify management personnel through channels such as phone calls, text messages, and mobile phone apps to ensure that the problem can be responded to and handled in a timely manner; S43: The system has an adaptive adjustment function. According to environmental changes and monitoring results, it automatically optimizes and adjusts detection parameters, such as acquisition frequency, sensitivity, etc., to improve monitoring accuracy and energy saving effects. When a device fails or the monitoring results are abnormal, the system will automatically detect and generate a detailed fault report, and then upload the report to the management platform through the remote maintenance module. After receiving the equipment fault report, the management platform can remotely diagnose the problem and perform maintenance.

Images