Rv1126-based tower online monitoring intelligent identification device

Abstract

The application discloses a kind of tower online monitoring intelligent identification device based on rv1126, comprising: image acquisition module, sensor data acquisition module, power data acquisition module, weather time obtaining module, weight distribution module and event probability solving module.The present application obtains the confidence of two similar events on image, sensor data and power data by image acquisition module, sensor data acquisition module, power data acquisition module;Again by weather event obtaining module and weight distribution module, obtain the influence of weather event to image data and sensor data respectively, adjust the weight of image data and sensor data in real time.Event probability solving module, according to each data and corresponding weight, obtains the occurrence probability of two similar events.The present application improves the accuracy of two similar tower event monitoring identification.

Description

Technical Field

[0001] This invention relates to the field of power equipment, and in particular to an intelligent identification device for online monitoring of power poles based on RV1126. Background Technology

[0002] Power transmission lines are supported by towers, which ensure that the distances between them and between them and the ground, as well as between them and the towers themselves, meet the requirements for electrical insulation safety and power frequency electromagnetic field limitation under various possible atmospheric conditions. Therefore, the safety of power transmission lines is a key issue in effectively guaranteeing power transmission. However, due to the special geographical location and environmental conditions of power transmission lines, they are numerous, widespread, and long, exposed to the elements year-round, and frequently subjected to severe weather events such as thunderstorms, strong winds, tornadoes, and hail. In addition to the natural environmental factors mentioned above, non-natural environmental factors may also have an impact on power transmission lines. These non-natural environmental factors include human factors and other non-human factors. Human factors include the reckless construction activities of cranes, excavators, and tower cranes near the power transmission lines. Non-human factors include flying objects (e.g., birds, unattended kites, sky lanterns, etc.) around the power transmission lines. Therefore, pole monitoring is a key aspect of power grid operation and maintenance. Timely detection and repair of pole failures can ensure the safety of core facilities such as high-voltage transmission towers and substation structures.

[0003] Current methods for monitoring power poles typically employ surveillance equipment, using images to obtain the pole's status. However, image recognition alone can be ineffective in distinguishing between various phenomena, such as wildfire smoke and fog, reflections from foreign objects, and electric arcs. Furthermore, the accuracy of image recognition is highly dependent on image quality. Extreme weather conditions or insufficient lighting in the area where the pole is located can easily affect image quality, leading to recognition errors and consequently impacting the accuracy of power pole monitoring and identification. Summary of the Invention

[0004] In view of the aforementioned deficiencies of the prior art, the technical problem to be solved by the present invention is to provide an intelligent identification device for online monitoring of poles and towers based on RV1126, which aims to improve the accuracy of monitoring and identification of two similar pole and tower events.

[0005] To achieve the above objectives, the present invention provides an intelligent identification device for online monitoring of power poles based on RV1126. The intelligent identification device for online monitoring of power poles includes: an image acquisition module, a sensor data acquisition module, a power data acquisition module, a weather time acquisition module, a weight allocation module, and an event probability solution module.

[0006] The image acquisition module is used to acquire monitoring images of a first monitoring area; identify the monitoring images to obtain a first similar event and its first image confidence level, and a second similar event and its second image confidence level corresponding to the monitoring images; wherein the images of the first similar event and the second similar event are similar in appearance;

[0007] The sensor data acquisition module is used to acquire sensor data from each sensor in the first monitoring area; and to generate a first sensor confidence score corresponding to the first similar event and a second sensor confidence score for the second similar event based on the sensor data.

[0008] The power data acquisition module is used to acquire power data of the first monitoring area; and to generate a first power confidence level corresponding to the first similar event and a second power confidence level of the second similar event based on the power data.

[0009] The weather and time acquisition module is used to acquire the current weather and time data of the first monitoring area;

[0010] The weight allocation module is used to obtain the image weight corresponding to the image acquisition module and the sensor weight corresponding to the sensor data acquisition module based on the weather data and the time data.

[0011] The event probability calculation module is used to obtain a first probability of occurrence of the first similar event by weighting the first image confidence and its corresponding image weight, the first sensor confidence and its corresponding sensor weight, and the first power confidence and its corresponding power weight; and to obtain a second probability of occurrence of the second similar event by weighting the second image confidence and its corresponding image weight, the second sensor confidence and its corresponding sensor weight, and the second power confidence and its corresponding power weight.

[0012] Optionally, the intelligent identification device for online monitoring of towers further includes: an event identification module;

[0013] The event recognition module is configured to, in response to the first occurrence probability being greater than the second occurrence probability, determine the first similar event as an event in progress and the second event as a misidentified event; or in response to the second occurrence probability being greater than the first occurrence probability, determine the second similar event as an event in progress and the first event as a misidentified event.

[0014] Optionally, the sensor corresponding to the sensor data acquisition module includes at least one of a temperature sensor, a humidity sensor, an ultraviolet sensor, and a PM value sensor.

[0015] Optionally, the weight allocation module is specifically used for:

[0016] Based on the weather data and the time data, native ambient light data is obtained; based on the native ambient light data, the image weight corresponding to the image acquisition module is determined; wherein, the larger the native ambient light data, the larger the image weight;

[0017] Based on the weather data and the time data, the native environmental values ​​corresponding to the detection values ​​of various sensors are obtained; based on the degree of influence of the native environmental values ​​on the corresponding sensors, the sensor weights corresponding to the sensor data acquisition modules are determined; wherein, the greater the degree of influence of the native environmental values ​​on the corresponding sensors, the smaller the sensor weights, and the sum of the image weights and the sensor weights is a constant value.

[0018] Optionally, the intelligent identification device for online monitoring of towers may further include: a shooting and adjustment module;

[0019] The shooting adjustment module is used to increase the aperture of the camera corresponding to the image acquisition module in response to the native ambient light data being less than a threshold, so as to increase the amount of light entering the camera.

[0020] Optionally, the weather time acquisition module is connected to an external server, specifically for:

[0021] By sending a weather and time query request to the external server, and receiving the current weather and time data of the first monitoring area corresponding to the query request from the external server.

[0022] Optionally, the power data includes at least one of current, voltage, power, electrical pulse frequency, high-frequency current distortion, and leakage current.

[0023] Optionally, the image acquisition module is further configured to:

[0024] In response to the fact that there is only one corresponding recognition event for the monitored image, the recognition time is determined to be the event that is currently occurring.

[0025] Optionally, the intelligent identification device for online monitoring of towers further includes: an event processing module;

[0026] The event processing module is used to determine whether the first similar event and the second similar event are fault hazard events; when the probability of occurrence of the fault hazard event is greater than that of the other event, a corresponding decision processing scheme is made according to the type of fault hazard event.

[0027] The beneficial effects of this invention are as follows: 1. This invention uses images, sensors, and power to monitor and identify similar events sequentially, obtaining data from all three sources. By cross-referencing this data, similar events are distinguished, thus improving the accuracy of monitoring and identification. Compared to relying solely on image recognition or image recognition combined with sensors, which is easily affected by external interference, this invention incorporates more data and makes judgments from more perspectives, improving the reliability of the identification results. 2. Based on current weather and time data, this invention analyzes the impact of environmental factors on image data acquisition and sensor data acquisition, and adjusts the confidence weights of both in real time according to the degree of impact, making the weighted sum of the two monitoring and identification results more reliable. Compared to directly fixing the confidence weights, adaptively adjusting the confidence weights based on environmental factors is more adaptable to different scenarios, improving the accuracy of monitoring and identification.

[0028] In summary, this invention improves the accuracy of monitoring and identifying two similar tower events by monitoring towers from different aspects, processing and adjusting the data in real time, and weighting and summing the data. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the structure of an intelligent identification device for online monitoring of power poles based on RV1126, provided in a specific embodiment of the present invention. Detailed Implementation

[0030] This invention discloses an intelligent identification device for online monitoring of power poles based on the RV1126. Those skilled in the art can refer to the content of this document and appropriately improve the technical details to implement it. It should be particularly noted that all similar substitutions and modifications are obvious to those skilled in the art and are considered to be included in this invention. The methods and applications of this invention have been described through preferred embodiments. Those skilled in the art can obviously modify or appropriately change and combine the methods and applications described herein without departing from the content, spirit, and scope of this invention to implement and apply the technology of this invention.

[0031] The applicant's research revealed that current monitoring of power poles typically utilizes surveillance equipment to obtain pole status information through images. However, image recognition alone can be ineffective in distinguishing between various phenomena, such as wildfire smoke and fog, reflections from foreign objects, and electric arcs. Furthermore, image recognition accuracy is highly dependent on image quality; extreme weather conditions or insufficient light in the pole's location, especially during peak hours, can easily affect image quality and lead to errors. Sensors can be used to detect various environmental parameters to reduce reliance on image recognition. For example, if wildfire smoke and fog cannot be distinguished, a humidity sensor can be used. However, sensors are also susceptible to external interference; in humid weather, it is difficult to differentiate between them using only sensors. Therefore, properly weighting image monitoring and sensor monitoring to adapt to changing environments is a crucial step for accurate identification.

[0032] Therefore, embodiments of the present invention provide an intelligent identification device for online monitoring of power poles based on RV1126, such as... Figure 1 As shown, the intelligent identification device for online monitoring of power poles includes: an image acquisition module 101, a sensor data acquisition module 102, a power data acquisition module 103, a weather time acquisition module 104, a weight allocation module 105, and an event probability solution module 106.

[0033] The image acquisition module 101 is used to acquire monitoring images of the first monitoring area; to identify the monitoring images and obtain the first similar event corresponding to the monitoring image and its first image confidence level, and the second similar event and its second image confidence level; wherein the images of the first similar event and the second similar event are similar in appearance;

[0034] The sensor data acquisition module 102 is used to acquire sensor data from various sensors in the first monitoring area; and to generate a first sensor confidence score corresponding to a first similar event and a second sensor confidence score for a second similar event based on the sensor data.

[0035] The power data acquisition module 103 is used to acquire power data in the first monitoring area; and based on the power data, to generate a first power confidence level corresponding to a first similar event and a second power confidence level for a second similar event.

[0036] Weather and time acquisition module 104 is used to acquire the current weather and time data of the first monitoring area;

[0037] The weight allocation module 105 is used to obtain the image weight corresponding to the image acquisition module 101 and the sensor weight corresponding to the sensor data acquisition module 102 based on weather data and time data.

[0038] The event probability calculation module 106 is used to obtain the first probability of occurrence of the first similar event by weighting the first image confidence and its corresponding image weight, the first sensor confidence and its corresponding sensor weight, and the first power confidence and its corresponding power weight; and to obtain the second probability of occurrence of the second similar event by weighting the second image confidence and its corresponding image weight, the second sensor confidence and its corresponding sensor weight, and the second power confidence and its corresponding power weight.

[0039] This invention identifies similar events by cross-referencing image data, sensor data, and power data, which can effectively improve the accuracy of identification.

[0040] In a specific application, such as wildfire smoke and morning fog, images represent two similar events. Even if they appear similar in an image, they can still differ, thus requiring different image confidence levels for two distinct similar events.

[0041] Image data: The smoke rises and spreads, and the morning fog is evenly distributed horizontally.

[0042] Sensor data: Smoke accompanied by a sudden temperature rise (>10℃ / min) and an increase in CO2 concentration. Morning fog had normal temperature and normal CO2 concentration.

[0043] Power data: Wildfires caused air ionization, leading to an increase in the frequency of corona discharge pulses on power lines. Corona discharge pulse frequencies were normal during the morning fog.

[0044] In the second specific application, metallic foreign object reflection and electric arc light are two similar events in the image data.

[0045] Image data: The reflective areas are bright but stable, and the arc light flickers and is accompanied by plasma glow.

[0046] Sensor data: Arc light generates ultraviolet radiation (UV sensor count > 500 times / s). Reflection from metallic foreign objects is normal.

[0047] Electrical data: Arc light causes high-frequency current distortion (sudden increase in 2-150kHz components). Reflection from metallic foreign objects is normal.

[0048] In the third specific application, insulator damage and contamination appear as two similar events in the image data.

[0049] Image data: The damaged edges are irregular, and the stains are distributed in patches.

[0050] Sensor data: Partial discharge in damaged insulator (UHF signal 300-1500MHz). Stains are normal.

[0051] Electrical data: Leakage current increased (from μA to mA). Contamination leakage current remained normal.

[0052] It should be noted that the RV1126 chip used in this invention possesses powerful image processing and AI inference capabilities, making it highly suitable for real-time monitoring systems. It offers advantages such as efficient image processing, AI inference capabilities, multi-interface scalability, and low-power design. By integrating the RV1126 end-side chip into the camera side of the transmission line tower, intelligent identification is achieved, primarily targeting hazards such as foreign objects in the conductors, construction machinery, smoke, and wildfires.

[0053] Furthermore, this invention not only identifies similar events using data from multiple sources, but also applies weight parameters to different aspects of the data based on environmental conditions, making the identification results more accurate. For example, on cloudy days or in the early morning or late evening when lighting conditions are poor, the images collected are not clear enough, so the weight of the image data can be appropriately reduced. Similarly, in humid weather such as rainy days, the humidity sensor is greatly affected by changes in similar events and environmental factors, so the weight of the humidity sensor can be appropriately reduced.

[0054] In this specific embodiment, the intelligent identification device for online monitoring of power poles further includes: an event identification module;

[0055] The event recognition module is used to determine the first similar event as an event in progress and the second event as a misidentified event in response to a first occurrence probability being greater than a second occurrence probability; or to determine the second similar event as an event in progress and the first event as a misidentified event in response to a second occurrence probability being greater than a first occurrence probability.

[0056] It should be noted that, under normal circumstances, the difference between the first and second occurrence probabilities can be quite large, because even if they appear similar in the image, they may differ significantly in other aspects, hence the large difference in their probabilities.

[0057] In this specific embodiment, the sensor corresponding to the sensor data acquisition module 102 includes at least one of a temperature sensor, a humidity sensor, an ultraviolet sensor, and a PM value sensor.

[0058] In this specific embodiment, the weight allocation module 105 is specifically used for:

[0059] Based on weather and time data, native ambient light data is obtained; based on the native ambient light data, the image weight corresponding to image acquisition module 101 is determined; where, the larger the native ambient light data, the larger the image weight.

[0060] Based on weather and time data, the native environmental values ​​of the corresponding detection values ​​of various sensors are obtained; based on the degree of influence of the native environmental values ​​on their corresponding sensors, the sensor weights corresponding to the sensor data acquisition module 102 are determined.

[0061] The greater the influence of the native environment value on its corresponding sensor, the smaller the sensor weight. The sum of the image weight and the sensor weight is a constant.

[0062] Native light data represents ambient light intensity that is unaffected by other factors.

[0063] In real-world scenarios, environmental factors often reduce the reliability of image or sensor data. For example, low ambient light can result in unclear images, leading to lower reliability. Therefore, it's necessary to adjust the confidence weights in real-time based on environmental data for better recognition.

[0064] It should be noted that the weighting is dynamic. When environmental factors have a greater impact on the image than on the sensor, the weight of sensor data is increased, and the weight of image data is decreased. Conversely, when environmental factors have a smaller impact on the sensor than on the image, the weight of image data is increased, and the weight of sensor data is decreased.

[0065] In this specific embodiment, the intelligent identification device for online monitoring of towers further includes: a shooting and adjustment module;

[0066] The shooting adjustment module is used to increase the aperture of the camera corresponding to the image acquisition module 101 in response to the original ambient light data being less than a threshold, so as to increase the amount of light entering the camera.

[0067] It should be noted that this module can make the images subsequently acquired by the image acquisition module 101 clearer under low light conditions.

[0068] In this specific embodiment, the weather time acquisition module 104 is connected to an external server, specifically for:

[0069] By sending a weather and time query request to an external server, and receiving the current weather and time data of the first monitoring area corresponding to the query request from the external server.

[0070] It should be noted that obtaining weather information through the internet requires no additional observation modules and is quite accurate.

[0071] In this specific embodiment, the image acquisition module 101 is further configured to:

[0072] In response to the fact that there is only one corresponding recognition event for the monitored image, the recognition time is determined as the event that is currently occurring.

[0073] It should be noted that most events are uniquely identified. Only some events have similarities and may be misidentified by image recognition. Therefore, when an event is unique, it can be confirmed that the event is happening.

[0074] In this specific embodiment, the intelligent identification device for online monitoring of towers further includes: an event processing module;

[0075] The event handling module is used to determine whether the first similar event and the second similar event are fault-hazard events; when the probability of occurrence of a fault-hazard event is greater than that of the other event, a corresponding decision-making and handling scheme is made according to the type of fault-hazard event.

[0076] It should be noted that similar events may both be hazard events, or one may be a hazard event while the other is not. When the probability of a hazard event occurring is greater than that of the other event, corresponding decision-making and handling plans need to be made to stop losses in time and prevent the hazard from continuing and causing further damage.

[0077] This invention employs image, sensor, and power monitoring and identification sequentially for similar events, obtaining data from all three sources. By cross-referencing this data, similar events are distinguished, thus improving the accuracy of similar event monitoring and identification. Compared to relying solely on image recognition or image recognition combined with sensors, which is susceptible to external interference, this invention incorporates more data and makes judgments from multiple perspectives, thereby enhancing the reliability of the identification results.

[0078] This invention analyzes the impact of environmental factors on image data acquisition and sensor data acquisition based on current weather and time data. It then adjusts the confidence weights of both data in real time according to the degree of impact, making the weighted sum of the two monitoring and identification results more reliable. Compared to directly fixing the confidence weights, adaptively adjusting the confidence weights based on environmental factors is more adaptable to different scenarios and improves the accuracy of monitoring and identification.

[0079] In summary, the embodiments of the present invention improve the accuracy of monitoring and identifying two similar tower events by monitoring towers from different aspects, processing and adjusting the data in real time, and weighting and summing the data.

[0080] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0081] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0082] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. An rv1126-based tower online monitoring intelligent identification device, characterized in that, The tower online monitoring intelligent identification device comprises an image acquisition module, a sensor data acquisition module, a power data acquisition module, a weather time acquisition module, a weight distribution module and an event probability solving module. The image acquisition module is configured to acquire a monitoring image of a first monitoring area, identify the monitoring image, and obtain a first similar event corresponding to the monitoring image and a first image confidence and a second similar event and a second image confidence; wherein the images of the first similar event and the second similar event are similar. The sensor data acquisition module is configured to acquire various sensor data of the first monitoring area, and generate a first sensor confidence corresponding to the first similar event and a second sensor confidence of the second similar event according to the sensor data. The power data acquisition module is configured to acquire power data of the first monitoring area, and generate a first power confidence corresponding to the first similar event and a second power confidence of the second similar event according to the power data. The weather time acquisition module is configured to acquire current weather data and time data of the first monitoring area. The weight distribution module is configured to obtain an image weight corresponding to the image acquisition module and a sensor weight corresponding to the sensor data acquisition module according to the weather data and the time data. The event probability solving module is configured to obtain a first occurrence probability of the first similar event by weighting according to the first image confidence and the corresponding image weight, the first sensor confidence and the corresponding sensor weight, and the first power confidence and the corresponding power weight; and obtain a second occurrence probability of the second similar event by weighting according to the second image confidence and the corresponding image weight, the second sensor confidence and the corresponding sensor weight, and the second power confidence and the corresponding power weight. The tower online monitoring intelligent identification device further comprises an event identification module; the event identification module is configured to determine the first similar event as an event occurring and the second similar event as a misrecognition event in response to the first occurrence probability being greater than the second occurrence probability, or determine the second similar event as an event occurring and the first similar event as a misrecognition event in response to the second occurrence probability being greater than the first occurrence probability. The weight distribution module is specifically configured to: obtain original ambient light data according to the weather data and the time data; determine the image weight corresponding to the image acquisition module according to the original ambient light data; wherein the greater the original ambient light data, the greater the image weight; obtain original ambient values of detection values corresponding to various sensors according to the weather data and the time data; determine the sensor weight corresponding to the sensor data acquisition module according to the influence degree of the original ambient value on the corresponding sensor; wherein the greater the influence degree of the original ambient value on the corresponding sensor, the smaller the sensor weight, and the sum of the image weight and the sensor weight is a constant value; The tower online monitoring intelligent identification device further comprises an event processing module; the event processing module is configured to determine whether the first similar event and the second similar event are fault hazard events; when the occurrence probability of the fault hazard event is greater than that of another event, a corresponding decision processing scheme is made according to the type of the fault hazard event.

2. The rv1126-based tower online monitoring intelligent identification device according to claim 1, characterized in that, The sensor corresponding to the sensor data acquisition module comprises at least one of a temperature sensor, a humidity sensor, an ultraviolet sensor and a PM value sensor.

3. The rv1126-based tower online monitoring intelligent identification device according to claim 1, characterized in that, The tower online monitoring intelligent identification device further comprises a shooting adjustment module. The shooting adjustment module is configured to increase the aperture of the camera corresponding to the image acquisition module to increase the light amount in response to the original ambient light data being less than a threshold value.

4. The rv1126-based tower online monitoring intelligent identification device according to claim 1, characterized in that, The weather time obtaining module is connected with an external server and is specifically configured to: send a weather time query request to the external server, and receive the current weather data and time data of the first monitoring area corresponding to the query request issued by the external server.

5. The rv1126-based tower online monitoring intelligent identification device according to claim 1, characterized in that, The power data comprises at least one of current, voltage, power, electric pulse frequency, high-frequency current distortion and leakage current.

6. The rv1126-based tower online monitoring intelligent identification device according to claim 1, characterized in that, The image acquisition module is further configured to: determine the recognition event as an event occurring in response to the monitoring image having only one corresponding recognition event.

Images