A transmission line project acceptance method and system based on drone

Abstract

The present application discloses a transmission line project acceptance method and system based on drones, which relates to the field of transmission project acceptance technology. The method includes: dividing the transmission line into different spatial regions, setting different acceptance qualification levels for different spatial regions; obtaining and analyzing component image information to obtain preliminary recognition results; performing abnormal credibility scoring on the preliminary recognition results, triggering reverse flight path signals, collecting and analyzing component reverse path images, obtaining reverse recognition results, judging the consistency between the preliminary recognition results and the reverse recognition results, and confirming that the component status is abnormal; counting the abnormal density of component status in each spatial region, obtaining component abnormal hot zones, and raising the acceptance qualification level of component abnormal hot zones. The present application realizes high-precision identification of the status of transmission line project components and regionalized risk response, thereby significantly improving the intelligence level and reliability of acceptance work.

Description

Technical Field

[0001] The present application relates to the technical field of power transmission project acceptance, and in particular to a method and system for power transmission line project acceptance based on drones. Background Art

[0002] Transmission projects are the "arteries" of the power system. Through high-voltage, high-capacity transmission networks, they optimize cross-regional energy allocation and support modern society's demand for high-reliability, high-efficiency, and low-carbon electricity. In power transmission line construction, completion acceptance is a critical step in ensuring line quality and safe operation. Through systematic testing and evaluation, we ensure the safe, stable, and efficient operation of the project, providing a guarantee for reliable power supply to the power system.

[0003] In related technologies, existing drone inspection systems generally use pre-set routes for one-way component acquisition. This is susceptible to light obstruction, angular distortion, or component orientation restrictions, resulting in incomplete image information, which in turn affects the accuracy of subsequent intelligent recognition and increases the risk of misjudgment or omission of abnormalities. Furthermore, traditional acceptance schemes often rely on manual intervention or fixed strategies after abnormal results appear. They lack intelligent feedback mechanisms based on risk density and are unable to upgrade the level or enhance image quality in high-risk areas. This leads to uneven resource allocation and low review efficiency, leaving room for improvement. Summary of the Invention

[0004] The purpose of the present invention is to provide a transmission line project acceptance method and system based on drones to solve the problems raised in the above background technology.

[0005] In the first aspect, the present application provides a transmission line project acceptance method based on drones, which adopts the following technical solutions:

[0006] Obtain satellite remote sensing data of the transmission line project, divide the transmission line into different spatial regions, and set different acceptance levels for different spatial regions;

[0007] Control the drone according to the set flight path to collect images of various engineering components on the transmission line and obtain component image information;

[0008] Performing recognition analysis on the component image information to obtain a preliminary recognition result, and performing an abnormality credibility score on the preliminary recognition result to obtain an abnormality credibility value;

[0009] Based on the abnormal credibility value, triggering a reverse flight path signal to control the drone to collect a reverse path image of the component;

[0010] Performing recognition analysis on the reverse path image to obtain a reverse recognition result, determining the consistency between the preliminary recognition result and the reverse recognition result, and confirming that the component state is abnormal;

[0011] The abnormal density of component states in the transmission lines in each spatial area is counted to obtain component abnormal hot zones, and the acceptance qualification level of the component abnormal hot zones is increased.

[0012] Preferably, the steps of obtaining satellite remote sensing data of the transmission line project, dividing the transmission line into different spatial regions, and setting different acceptance levels for the different spatial regions are specifically as follows:

[0013] Acquiring satellite remote sensing data covering the power transmission line project, wherein the satellite remote sensing data includes the path of the power transmission line and its geographical environment information;

[0014] Based on the transmission line path, dividing the transmission line into a plurality of different spatial regions;

[0015] Based on the geographic environment information, environmental characteristic information of each spatial area is extracted, and corresponding acceptance grades are set for different spatial areas according to the environmental characteristic information.

[0016] Preferably, a UAV flight path is formed to obtain spatial distribution information of transmission line engineering components, a one-to-one correspondence is established between waypoints in the UAV flight path and various engineering components on the transmission line, and a component image acquisition list is formed;

[0017] When the UAV executes the flight path, the distance parameter between the UAV's current position and the target component is obtained;

[0018] Based on the distance parameter, the drone is controlled to automatically capture an image of the target component to obtain component image information.

[0019] Preferably, based on the distance parameter, the step of controlling the drone to automatically capture the target component image and obtaining component image information is specifically as follows:

[0020] Setting a component image acquisition distance threshold, and comparing the distance parameter with the component image acquisition distance threshold;

[0021] If the distance parameter is less than the component image acquisition distance threshold, the angle parameter between the current heading of the UAV and the orientation of the target component is obtained;

[0022] Setting a component image acquisition angle deviation range, and comparing the angle parameter with a component image acquisition angle threshold;

[0023] If the angle parameter is within the component image acquisition angle deviation range, an image acquisition signal is triggered to control the UAV to automatically capture the target component image to obtain component image information.

[0024] Preferably, the steps of identifying and analyzing the component image information, calculating the component abnormality credibility score, and obtaining the component abnormality credibility value are specifically as follows:

[0025] Based on the component image information, extract component type information, structural feature information, and appearance state information to generate a preliminary recognition result;

[0026] Extracting clarity parameters and occlusion rate parameters of the component image based on the preliminary recognition result, and combining the clarity parameters and occlusion rate parameters to obtain a component image quality score;

[0027] Based on the preliminary recognition results, extracting structural information and state information of the current component and the adjacent components before and after, and combining the structural information and state information to obtain a consistency score of the adjacent components;

[0028] Based on the component image quality score and the adjacent component consistency score, a component anomaly credibility score is obtained by weighted synthesis to generate an anomaly credibility value.

[0029] Preferably, based on the abnormal credibility value, the step of triggering the reverse flight path signal and controlling the drone to collect the reverse path image of the component is specifically as follows:

[0030] Based on the abnormality credibility value, setting an abnormality credibility threshold, and comparing the abnormality credibility value with the abnormality credibility threshold;

[0031] If the abnormality credibility value is higher than the abnormality credibility threshold, it is determined that a high-credibility abnormality exists in the current component, and a reverse flight path signal is triggered;

[0032] Based on the reverse flight path signal, the UAV is controlled to generate a reverse flight path of the current component and collect a reverse path image of the current component.

[0033] Preferably, the steps of performing recognition analysis on the reverse path image to obtain a reverse recognition result, judging the consistency between the preliminary recognition result and the reverse recognition result, and confirming that the component state is abnormal are specifically as follows:

[0034] Performing recognition analysis on the reverse path image, extracting component type information, structural feature information, and appearance status information, and generating a reverse recognition result;

[0035] Extracting the preliminary recognition result, comparing the preliminary recognition result with the reverse recognition result for consistency, and determining whether the two recognition results are consistent;

[0036] If the preliminary recognition result is inconsistent with the reverse recognition result, it will be marked as a recognition conflict and enter the manual review process;

[0037] If the preliminary identification result is consistent with the reverse identification result, it is confirmed that the component state is abnormal.

[0038] Preferably, the steps of counting the abnormal density of component states in the transmission line in each spatial region to obtain abnormal component hot zones and raising the acceptance grade of the abnormal component hot zones are specifically as follows:

[0039] Count the number of abnormal status components in the transmission lines in each spatial area and the total number of components in the area;

[0040] Calculating the ratio of the number of components with abnormal status to the total number of components in the region to obtain the abnormal density of components in the transmission line in each spatial region;

[0041] Setting an abnormal density threshold, and comparing the abnormal density of the component with the abnormal density threshold;

[0042] If the component abnormality density exceeds the abnormality density threshold, the spatial area is determined to be a component abnormality hot zone;

[0043] The acceptance level of the abnormal hot zone of the component is raised, the threshold of the acquired image quality is lowered and the error range of the shooting angle is narrowed.

[0044] Secondly, this application provides a UAV-based transmission line project acceptance system that adopts the following technical solutions:

[0045] A transmission line project acceptance system based on drones, comprising:

[0046] The acceptance level setting module obtains satellite remote sensing data of the transmission line project, divides the transmission line into different spatial areas, and sets different acceptance levels for different spatial areas;

[0047] Build an image acquisition module to control the drone according to the set flight path to collect images of various engineering components on the transmission line and obtain component image information;

[0048] An abnormal credibility scoring module performs recognition analysis on the component image information to obtain a preliminary recognition result, and performs abnormal credibility scoring on the preliminary recognition result to obtain an abnormal credibility value;

[0049] a reverse path triggering module, which triggers a reverse flight path signal based on the abnormal credibility value and controls the UAV to collect a reverse path image of the component;

[0050] Constructing an abnormality confirmation module to perform recognition analysis on the reverse path image to obtain a reverse recognition result, judging the consistency between the preliminary recognition result and the reverse recognition result, and confirming that the component state is abnormal;

[0051] The acceptance level adjustment module counts the abnormal density of component states in the transmission line in each spatial area, obtains component abnormal hot zones, and increases the acceptance level of the component abnormal hot zones.

[0052] In summary, this application includes at least one of the following beneficial technical effects:

[0053] 1. By dividing the line into multiple spatial zones based on environmental characteristics, differentiated acceptance levels can be set based on regional characteristics. This improves the adaptability and precision of the acceptance standards, avoids a one-size-fits-all approach, and enhances overall acceptance efficiency and risk control capabilities. Image processing and feature recognition technologies extract component-related feature information. Combined with anomaly confidence scoring rules, a component anomaly confidence value is calculated to determine whether the component is likely to be an anomaly. If the initial component identification result is insufficiently reliable, a reverse path acquisition mechanism is dynamically triggered, allowing the drone to capture the same component again from different angles or paths. This compensates for blind spots in single-view recognition, eliminates misjudgments caused by lighting, angle, and occlusion, and improves system robustness and accuracy. By comparing the forward and reverse path recognition results, component status is further confirmed. If the two recognitions agree, the component is directly identified as an anomaly, significantly improving recognition accuracy. If the results conflict, manual review is performed to avoid misjudgments caused by single-time recognition errors, ensuring the reliability of the acceptance results while reducing the risks and potential risks of resuming work due to misjudgments. By identifying hot spots with abnormal density of local components, we can perceive and respond to regional risks. After determining the area with high abnormal density as an abnormal hot zone, we will automatically increase the acceptance level of the components in this area to ensure that high-risk areas are inspected more strictly, prevent risk spillover, improve the accuracy and controllability of the acceptance system, and upgrade the acceptance work from single-point intelligence to an intelligent mode of regional perception and level self-adjustment.

[0054] 2. Build an intelligent judgment trigger mechanism to compare the abnormal credibility value of the component identified by the system with the preset credibility threshold. If the abnormal credibility value exceeds the set threshold, it can be determined that the abnormal state of the component is highly credible. At this time, the system automatically generates and triggers the reverse flight path signal to provide control instructions for subsequent review work. In this way, the system establishes an intelligent review trigger mechanism, avoiding the inflexibility of manually setting fixed review areas or time nodes, making the image re-collection behavior more targeted and timely, and greatly improving the accuracy and credibility of abnormality identification. Based on the triggered reverse flight path signal, the drone automatically replans the flight path, allowing it to re-take images of the current component from relative angles, opposite directions or different postures. Compared with a single shooting angle, the reverse path image can provide more details, different lighting and structural surfaces, which helps to discover hidden defects or eliminate misjudgments, make up for the blind spot problem caused by single-angle shooting, and effectively improve the accuracy of image recognition.

[0055] 3. By identifying and counting the number of components identified as abnormal in each spatial region, and simultaneously obtaining the total number of all components in the region, we calculate the abnormal component ratio (the ratio of the number of abnormal components to the total number of components in the region), quantifying the density of anomalies within a local area. Compared to single anomaly points, the anomaly density index better reflects regional risk trends and systemic quality issues, achieving a strategic transition from single-point precision to regional early warning. If the density of component anomalies in a certain spatial area exceeds the set threshold, the area is regarded as a component anomaly hot zone, which means that the area has high structural risks, construction quality fluctuations or environmental interference problems. By dynamically adjusting the acceptance standards, the image quality requirements and anomaly judgment threshold of the components in the area are improved, that is, the image quality threshold is raised to require images with higher clarity, more correct perspective, and less occlusion; the anomaly credibility threshold is raised to prevent low-confidence anomalies from being misjudged as problems, and a higher credibility is required to trigger a review, thereby strengthening the recognition ability and robustness in key risk areas, and preventing chain misjudgments or wrong repairs caused by abnormal density in hot zones; at the same time, it effectively saves manual review resources, and realizes an intelligent acceptance resource allocation model that raises standards when risks are high and relaxes standards when risks are low. This not only enhances the system's ability to identify and respond to high-risk areas, but also realizes the on-demand and graded dynamic evolution of acceptance standards. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram of the specific steps of an embodiment of a transmission line project acceptance method based on drones of the present invention.

[0057] Figure 2 This is a module connection diagram of an embodiment of a transmission line project acceptance system based on a drone of the present invention. DETAILED DESCRIPTION

[0058] Below is a combination of the embodiments and Figure 1-Figure 2 The present invention will be described in further detail, but the embodiments of the present invention are not limited thereto.

[0059] The present invention discloses a transmission line project acceptance method based on a drone, which specifically includes the following steps:

[0060] Step S1, obtaining satellite remote sensing data of the transmission line project, dividing the transmission line into different spatial regions, and setting different acceptance levels for different spatial regions;

[0061] Step S2: Control the UAV according to the set flight path to collect images of various engineering components on the transmission line to obtain component image information;

[0062] Step S3, performing recognition analysis on the component image information to obtain a preliminary recognition result, and performing an abnormality credibility score on the preliminary recognition result to obtain an abnormality credibility value;

[0063] Step S4: triggering a reverse flight path signal based on the abnormal credibility value to control the UAV to collect a reverse path image of the component;

[0064] Step S5, performing recognition analysis on the reverse path image to obtain a reverse recognition result, judging the consistency between the preliminary recognition result and the reverse recognition result, and confirming that the component state is abnormal;

[0065] Step S6: Count the abnormal density of component states in the power transmission line in each spatial area to obtain abnormal component hot zones, and increase the acceptance level of the abnormal component hot zones.

[0066] In practical applications, by dividing the line into multiple spatial zones based on environmental characteristics, differentiated acceptance criteria can be set based on regional characteristics. For example, higher image quality thresholds are recommended in mountainous areas with strong winds and significant obstruction, while higher anomaly confidence thresholds are required in urban areas. This improves the adaptability and precision of acceptance criteria, avoids a one-size-fits-all approach, and enhances overall acceptance efficiency and risk control capabilities. By combining regional levels with component distribution data, an optimal flight path is planned and drones are controlled to capture images of each component along the set route, achieving efficient and high-coverage component inspections. Image processing and feature recognition technologies extract component-related feature information and, combined with anomaly confidence scoring rules, calculate a component anomaly confidence value to determine whether the component is likely to be an anomaly. If the initial component recognition result is insufficiently reliable, a reverse path acquisition mechanism is dynamically triggered, causing the drone to capture the same component a second time from different angles or paths. This compensates for blind spots in single-view recognition, eliminates misjudgments due to lighting, angle, and obstruction, and improves system robustness and accuracy. By comparing the similarities and differences between the forward and reverse path recognition results, the component status can be further confirmed. If the two recognitions are consistent, it can be directly confirmed as an abnormal state, significantly improving the recognition accuracy. If the results conflict, manual review is performed to avoid misjudgment due to single recognition errors, providing reliability guarantees for the acceptance results, while reducing the risk of resumption of work and hidden dangers caused by misjudgment. By aggregating component abnormal status data in the spatial dimension, hot zone sections with dense and abnormal local components are identified, and regional risks are perceived and responded to. After determining an area with a high abnormal density as an abnormal hot zone, the acceptance level of the components in the area is automatically increased, such as increasing image clarity requirements, increasing the number of reviews, etc., to ensure that high-risk areas are more strictly inspected, prevent risk spillover, and improve the accuracy and controllability of the acceptance system, so that the acceptance work is upgraded from single-point intelligence to an intelligent mode of regional perception plus level self-adjustment.

[0067] The steps for obtaining satellite remote sensing data of the transmission line project, dividing the transmission line into different spatial regions, and setting different acceptance levels for different spatial regions are as follows:

[0068] Step S11, acquiring satellite remote sensing data covering the power transmission line project, wherein the satellite remote sensing data includes the power transmission line path and its geographical environment information;

[0069] Step S12, dividing the transmission line into a plurality of different spatial regions based on the transmission line path;

[0070] Step S13: extracting environmental characteristic information of each spatial area based on the geographic environment information, and setting corresponding acceptance levels for different spatial areas according to the environmental characteristic information.

[0071] In practical applications, satellite remote sensing data is used to obtain high-altitude, macroscopic images of the overall spatial distribution of transmission lines and their surroundings, compensating for the high costs, numerous blind spots, and slow coverage of traditional manual or ground-based surveying. Remote sensing data not only provides the trajectory of transmission lines but also contains rich geographic information, such as topography, river and mountain distribution, vegetation density, and land use types. This scientifically divides the entire transmission line project into several independent spatial units, slicing the previously continuous and complex line structure into independently assessed and dynamically managed acceptance units, facilitating the implementation of differentiated standards across diverse environments and risk profiles. By analyzing the environmental characteristic information corresponding to each spatial area, such as terrain height, occlusion risk, corrosive climate, vegetation cover rate, intensity of human interference, etc., the difficulty and risk of regional construction and component operation are quantitatively assessed, and corresponding acceptance standards are set according to different risk levels. For example, the abnormal credibility threshold requirements are increased in high-corrosion areas, and the image acquisition angle restrictions are increased in mountainous occlusion areas. This allows the acceptance intensity to be increased on demand, significantly improving the accuracy and risk adaptability of the acceptance system, avoiding waste of resources in low-risk areas and failure of acceptance in high-risk areas.

[0072] The steps for controlling the drone to collect images of various engineering components on the transmission line according to the set flight path and obtaining component image information are as follows:

[0073] Step S21: Generate a UAV flight path based on the transmission line path, obtain spatial distribution information of transmission line engineering components, establish a one-to-one correspondence between waypoints in the UAV flight path and each engineering component on the transmission line, and form a component image acquisition list;

[0074] Step S22, when the UAV executes the flight path, obtaining the distance parameter between the current position of the UAV and the target component;

[0075] Step S23: Based on the distance parameter, the drone is controlled to automatically capture an image of the target component to obtain component image information.

[0076] In actual application, the structural path information of the transmission line is integrated with the spatial positioning information of the engineering components to generate a drone route with high coverage and excellent flight efficiency. By laying out waypoints in the flight path and establishing a one-to-one mapping relationship between each waypoint and the corresponding component, the image acquisition position, angle and height of each component can be planned in advance to form a component image acquisition list. This effectively solves the problems of close proximity but missed shots or inconsistent shots in traditional manual inspections, and is a prerequisite for achieving high-precision component data acquisition. During the execution of the drone, its current GPS coordinates are continuously collected, compared with the known component coordinates in real time, and the distance between the current position and the target component is calculated. This ensures that the drone can sense whether the target is close to the acquisition area during flight, avoiding misjudgment or missed shots due to factors such as excessive speed and GPS drift, thereby improving the accuracy and stability of image capture. The distance-based automatic triggering method makes the shooting behavior more intelligent and accurate, ensuring the optimal image acquisition timing and reasonable viewing angle, effectively avoiding the occurrence of low-quality images such as blur, deflection, and occlusion. By obtaining standardized component image information, it provides a stable and reliable data foundation for subsequent AI recognition and status judgment, thereby improving the intelligence level and engineering controllability of the entire acceptance system.

[0077] Based on the distance parameter, the steps of controlling the drone to automatically capture the target component image and obtain component image information are specifically as follows:

[0078] Step S231, setting a component image acquisition distance threshold, and comparing the distance parameter with the component image acquisition distance threshold;

[0079] Step S232: If the distance parameter is less than the component image acquisition distance threshold, then obtain the angle parameter between the current heading of the UAV and the orientation of the target component;

[0080] Step S233, setting a component image acquisition angle deviation range, and comparing the angle parameter with a component image acquisition angle threshold;

[0081] Step S234: If the angle parameter is within the component image acquisition angle deviation range, an image acquisition signal is triggered to control the drone to automatically capture the target component image to obtain component image information.

[0082] In practical application, a preset component image acquisition distance threshold defines the effective distance range between the drone and the target component for image acquisition, ensuring that acquisition is performed only when the component is within the field of view. Comparing the real-time distance between the drone's current position and the component's location prevents premature capture from a distance or missed targets due to close proximity, ensuring that image acquisition is performed within a reasonable spatial range. The angle between the drone's flight direction and the component's spatial orientation is calculated to determine whether the drone is within the component's effective image acquisition angle range. Because the physical structure of components, such as crossarms, hardware, and insulators, exhibits significant directionality, clear and readable images can only be obtained from appropriate angles. This overcomes the limitations of distance estimation alone and ensures image quality in the angular dimension. Comparing the current angle with this threshold not only filters out invalid images with excessively large deviations but also accommodates angular perturbations caused by minor attitude changes during flight, thereby achieving robust control of image acquisition and enhancing the adaptive capabilities of the intelligent image acquisition system. When the drone meets both the distance and angle conditions, it automatically triggers an image acquisition signal, achieving standardized and precise capture of the target component. It effectively avoids problems such as image blur, occlusion, and distortion caused by angle offset and insufficient distance. The component image information finally generated has good clarity, angle visibility, and feature integrity, providing high-quality data support for subsequent image recognition and status judgment, and is a key guarantee for data reliability in the intelligent acceptance system.

[0083] The steps of identifying and analyzing the component image information, calculating the component abnormality credibility score, and obtaining the component abnormality credibility value are specifically as follows:

[0084] Step S31, based on the component image information, extract component type information, structural feature information and appearance state information to generate a preliminary recognition result;

[0085] Step S32: extracting a clarity parameter and an occlusion rate parameter of the component image based on the preliminary recognition result, and synthesizing the clarity parameter and the occlusion rate parameter to obtain a component image quality score;

[0086] Step S33: Based on the preliminary recognition result, extract the structural information and state information of the current component and the adjacent components before and after, and combine the structural information and state information to obtain the consistency score of the adjacent components;

[0087] Step S34: Based on the component image quality score and the adjacent component consistency score, a component abnormality credibility score is obtained by weighted synthesis to generate an abnormality credibility value.

[0088] In practical applications, image recognition processes captured component images to extract the target component's basic type, structural features (such as shape, outline, and connection type), and appearance (such as damage, corrosion, and offset), and to preliminarily determine whether the component is abnormal. Low-level visual features such as image clarity (such as edge sharpness and grayscale contrast) and occlusion (such as the proportion of occlusion by vegetation, foreign objects, or the component itself) are analyzed to quantify the image's recognizability. Combining these two key parameters, an image quality score is generated. If image quality is poor, even if the initial recognition result indicates an anomaly, its credibility should be lowered to prevent false positives and ensure the reliability of the recognition conclusion and project feasibility. Components in transmission lines are typically arranged in groups and continuously, and their structural properties and operating states exhibit distinct continuity. By extracting information such as the structural type, size ratio, and state recognition results of adjacent components, it is possible to determine whether the current component is within reasonable structural continuity. For example, if the state of a central component is significantly abnormal while the upper and lower components are normal, this may be an identification error. This can enhance the system's ability to identify isolated abnormal points, effectively suppressing identification interference such as single-point misjudgments and edge anomalies, and improving system stability. Through weighted structural fusion, the image quality score is combined with the adjacent consistency score to form a quantifiable anomaly confidence value, which fluctuates within the range of [0, 1] and is used to determine the confidence level of anomaly determinations. This effectively enhances the credibility assessment capabilities of transmission line component identification results, improving the accuracy and robustness of anomaly identification and providing precise triggering for subsequent reverse verification, hot spot determination, and level upgrades.

[0089] The steps of triggering a reverse flight path signal based on the abnormal credibility value and controlling the drone to collect a reverse path image of the component are specifically as follows:

[0090] Step S41, setting an abnormality credibility threshold based on the abnormality credibility value, and comparing the abnormality credibility value with the abnormality credibility threshold;

[0091] Step S42: If the abnormality credibility value is higher than the abnormality credibility threshold, it is determined that a high-credibility abnormality exists in the current component, and a reverse flight path signal is triggered;

[0092] Step S43: Based on the reverse flight path signal, the UAV is controlled to generate a reverse flight path of the current component and collect a reverse path image of the current component.

[0093] In actual application, an intelligent judgment trigger mechanism is constructed to compare the abnormal credibility value of the component identified by the system with the preset credibility threshold. If the abnormal credibility value exceeds the set threshold, it can be determined that the abnormal state of the component is highly credible. At this time, the system automatically generates and triggers the reverse flight path signal to provide control instructions for subsequent review work. In this way, the system establishes an intelligent review trigger mechanism, avoiding the inflexibility of manually setting fixed review areas or time nodes, making the image re-collection behavior more targeted and timely, and greatly improving the accuracy and credibility of abnormality identification. Based on the triggered reverse flight path signal, the drone automatically replans the flight path, allowing it to re-take images of the current component from relative angles, opposite directions or different postures. Compared with a single shooting angle, the reverse path image can provide more details, different lighting and structural surfaces, which helps to discover hidden defects or eliminate misjudgments, make up for the blind spot problem caused by single-angle shooting, and effectively improve the accuracy of image recognition.

[0094] The steps of performing recognition analysis on the reverse path image to obtain a reverse recognition result, judging the consistency between the preliminary recognition result and the reverse recognition result, and confirming that the component state is abnormal are specifically as follows:

[0095] Step S51, performing recognition analysis on the reverse path image, extracting component type information, structural feature information, and appearance state information, and generating a reverse recognition result;

[0096] Step S52: extract the preliminary recognition result, compare the preliminary recognition result with the reverse recognition result, and determine whether the two recognition results are consistent;

[0097] Step S53: If the preliminary recognition result is inconsistent with the reverse recognition result, it is marked as a recognition conflict and enters the manual review process;

[0098] Step S54: If the preliminary recognition result is consistent with the reverse recognition result, it is confirmed that the component state is abnormal.

[0099] In actual use, by re-extracting information such as the component type, structural features, and appearance status from the reverse image, the system can generate a reverse recognition result independent of the initial recognition path. By comparing structural features, identification labels, damage types, or location descriptions, the system can determine whether the two recognition results are consistent. Consistency comparison can not only identify abnormal components with overlapping confirmations, but also detect single recognition misjudgments caused by viewing angle problems, significantly improving the system's confidence in abnormal state judgments. When two recognition results conflict, for example, one is identified as rust and the other is normal, it will be marked as an identification conflict and a manual review process will be automatically introduced to prevent the system from mistakenly identifying certain boundary features. At the same time, it takes into account the efficiency of automation and the accuracy of manual judgment, realizing a double-layer protection of intelligent judgment and manual backup, which meets the high requirements of accuracy and reliability in engineering acceptance scenarios. When the forward and reverse recognition results are highly consistent, it means that the current component anomaly has been verified and supported from different perspectives. At this time, it can be confirmed that the status judgment of the abnormal component is valid, realizing a closed-loop intelligent judgment process from image acquisition to result confirmation, effectively improving the accuracy and credibility of recognition, and enhancing the stability of the system's judgment on component status under complex working conditions.

[0100] The steps of counting the abnormal density of component states in the transmission line in each spatial area to obtain component abnormal hot zones and raising the acceptance grade of the component abnormal hot zones are specifically as follows:

[0101] Step S61, counting the number of abnormal state components in the power transmission lines in each spatial area and the total number of components in the area;

[0102] Step S62, calculating the ratio of the number of abnormal components to the total number of components in the region, and obtaining the abnormal density of components in the transmission line in each spatial region;

[0103] Step S63, setting an abnormal density threshold, and comparing the abnormal density of the component with the abnormal density threshold;

[0104] Step S64: if the component abnormality density exceeds the abnormality density threshold, the spatial area is determined to be a component abnormality hot zone;

[0105] Step S65 , raising the acceptance level of the abnormal hot zone of the component, and increasing the captured image quality threshold and the abnormality credibility threshold.

[0106] In practical applications, by identifying and counting the number of components judged to be in an abnormal state in each spatial area, and simultaneously obtaining the total number of all components in the area, the abnormal component ratio indicator is calculated (that is, the ratio of the number of abnormal components to the total number of components in the area), quantifying the density of anomalies in a local area. Compared with the judgment of a single abnormal point, the abnormal density indicator can better reflect regional risk trends and systemic quality issues. For example, if multiple consecutive tower components in the same line section are abnormal, even if the credibility of each anomaly is not high, the abnormal density may reach the trigger threshold, thereby raising the acceptance standard and achieving a strategic transition from single-point precision to regional early warning. If the density of component anomalies in a certain spatial area exceeds the set threshold, the area is regarded as a component anomaly hot zone, which means that the area has high structural risks, construction quality fluctuations or environmental interference problems. By dynamically adjusting the acceptance standards, the image quality requirements and anomaly judgment threshold of the components in the area are improved, that is, the image quality threshold is raised to require images with higher clarity, more correct perspective, and less occlusion; the anomaly credibility threshold is raised to prevent low-confidence anomalies from being misjudged as problems, and a higher credibility is required to trigger a review, thereby strengthening the recognition ability and robustness in key risk areas, and preventing chain misjudgments or wrong repairs caused by abnormal density in hot zones; at the same time, it effectively saves manual review resources, and realizes an intelligent acceptance resource allocation model that raises standards when risks are high and relaxes standards when risks are low. This not only enhances the system's ability to identify and respond to high-risk areas, but also realizes the on-demand and graded dynamic evolution of acceptance standards.

[0107] A UAV-based power transmission line project acceptance system, which applies the above-mentioned UAV-based power transmission line project acceptance method, includes:

[0108] The acceptance level setting module obtains satellite remote sensing data of the transmission line project, divides the transmission line into different spatial areas, and sets different acceptance levels for different spatial areas;

[0109] Build an image acquisition module to control the drone according to the set flight path to collect images of various engineering components on the transmission line and obtain component image information;

[0110] An abnormal credibility scoring module performs recognition analysis on the component image information to obtain a preliminary recognition result, and performs abnormal credibility scoring on the preliminary recognition result to obtain an abnormal credibility value;

[0111] a reverse path triggering module, which triggers a reverse flight path signal based on the abnormal credibility value and controls the UAV to collect a reverse path image of the component;

[0112] Constructing an abnormality confirmation module to perform recognition analysis on the reverse path image to obtain a reverse recognition result, judging the consistency between the preliminary recognition result and the reverse recognition result, and confirming that the component state is abnormal;

[0113] The acceptance level adjustment module counts the abnormal density of component states in the transmission line in each spatial area, obtains component abnormal hot zones, and increases the acceptance level of the component abnormal hot zones.

[0114] In practical application, the acceptance level setting module divides transmission lines into distinct spatial regions based on satellite remote sensing data and assigns differentiated acceptance levels to each region, enabling zoning acceptance management. The image acquisition module controls drones along pre-set flight paths to capture images of transmission line components, ensuring complete, high-quality component visual information. The anomaly credibility scoring module then intelligently identifies and assesses the quality of the captured component images, generating preliminary identification results and calculating component anomaly credibility values ​​to assist in anomaly risk assessment. The reverse path triggering module then determines whether to trigger a reverse flight path based on the anomaly credibility values, controlling the drone to re-image the suspected component from the opposite angle to enhance identification reliability. The anomaly confirmation module then identifies the reverse path images and compares them with the preliminary results to confirm whether any component anomalies are present, improving assessment accuracy. Finally, the acceptance level adjustment module dynamically identifies anomaly hotspots based on the density of anomaly components within each spatial region and adjusts their acceptance levels upward, establishing an intelligent, risk-driven acceptance control mechanism.

[0115] The above are all preferred embodiments of the present application, and are not intended to limit the scope of protection of the present application. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A transmission line project acceptance method based on drones, characterized in that: The following steps are involved: Obtain satellite remote sensing data of the transmission line project, divide the transmission line into different spatial regions, and set different acceptance levels for different spatial regions; Control the drone according to the set flight path to collect images of various engineering components on the transmission line and obtain component image information; Performing recognition analysis on the component image information to obtain a preliminary recognition result, and performing an abnormality credibility score on the preliminary recognition result to obtain an abnormality credibility value; Based on the abnormal credibility value, triggering a reverse flight path signal to control the drone to collect a reverse path image of the component; Performing recognition analysis on the reverse path image to obtain a reverse recognition result, determining the consistency between the preliminary recognition result and the reverse recognition result, and confirming that the component state is abnormal; The abnormal density of component states in the transmission lines in each spatial area is counted to obtain component abnormal hot zones, and the acceptance qualification level of the component abnormal hot zones is increased.

2. The transmission line project acceptance method based on drone according to claim 1, characterized in that: The steps of obtaining satellite remote sensing data of the transmission line project, dividing the transmission line into different spatial regions, and setting different acceptance levels for different spatial regions are specifically as follows: Acquiring satellite remote sensing data covering the power transmission line project, wherein the satellite remote sensing data includes the path of the power transmission line and its geographical environment information; Based on the transmission line path, dividing the transmission line into a plurality of different spatial regions; Based on the geographic environment information, environmental characteristic information of each spatial area is extracted, and corresponding acceptance grades are set for different spatial areas according to the environmental characteristic information.

3. The transmission line project acceptance method based on drone according to claim 1 is characterized in that: The steps of controlling the drone according to the set flight path to collect images of various engineering components on the transmission line and obtain component image information are specifically as follows: Generate a UAV flight path based on the transmission line path, obtain the spatial distribution information of the transmission line engineering components, establish a one-to-one correspondence between the waypoints in the UAV flight path and each engineering component on the transmission line, and form a component image collection list; When the UAV executes the flight path, the distance parameter between the UAV's current position and the target component is obtained; Based on the distance parameter, the drone is controlled to automatically capture an image of the target component to obtain component image information.

4. The transmission line project acceptance method based on drone according to claim 3 is characterized in that: The step of controlling the drone to automatically capture the target component image based on the distance parameter to obtain component image information is specifically as follows: Setting a component image acquisition distance threshold, and comparing the distance parameter with the component image acquisition distance threshold; If the distance parameter is less than the component image acquisition distance threshold, the angle parameter between the current heading of the UAV and the orientation of the target component is obtained; Setting a component image acquisition angle deviation range, and comparing the angle parameter with a component image acquisition angle threshold; If the angle parameter is within the component image acquisition angle deviation range, an image acquisition signal is triggered to control the UAV to automatically capture the target component image to obtain component image information.

5. The transmission line project acceptance method based on drone according to claim 1 is characterized in that: The step of identifying and analyzing the component image information, calculating the component abnormality credibility score, and obtaining the component abnormality credibility value is specifically as follows: Based on the component image information, extract component type information, structural feature information, and appearance state information to generate a preliminary recognition result; Extracting clarity parameters and occlusion rate parameters of the component image based on the preliminary recognition result, and combining the clarity parameters and occlusion rate parameters to obtain a component image quality score; Based on the preliminary recognition results, extracting structural information and state information of the current component and the adjacent components before and after, and combining the structural information and state information to obtain a consistency score of the adjacent components; Based on the component image quality score and the adjacent component consistency score, a component anomaly credibility score is obtained by weighted synthesis to generate an anomaly credibility value.

6. The transmission line project acceptance method based on drone according to claim 1, characterized in that: The step of triggering a reverse flight path signal based on the abnormal credibility value and controlling the drone to collect a reverse path image of the component is specifically as follows: Based on the abnormality credibility value, setting an abnormality credibility threshold, and comparing the abnormality credibility value with the abnormality credibility threshold; If the abnormality credibility value is higher than the abnormality credibility threshold, it is determined that a high-credibility abnormality exists in the current component, and a reverse flight path signal is triggered; Based on the reverse flight path signal, the UAV is controlled to generate a reverse flight path of the current component and collect a reverse path image of the current component.

7. The transmission line project acceptance method based on drone according to claim 5, characterized in that: The steps of performing recognition analysis on the reverse path image to obtain a reverse recognition result, judging the consistency between the preliminary recognition result and the reverse recognition result, and confirming that the component state is abnormal are specifically as follows: Performing recognition analysis on the reverse path image, extracting component type information, structural feature information, and appearance status information, and generating a reverse recognition result; Extracting the preliminary recognition result, comparing the preliminary recognition result with the reverse recognition result for consistency, and determining whether the two recognition results are consistent; If the preliminary recognition result is inconsistent with the reverse recognition result, it will be marked as a recognition conflict and enter the manual review process; If the preliminary identification result is consistent with the reverse identification result, it is confirmed that the component state is abnormal.

8. The transmission line project acceptance method based on drone according to claim 1, characterized in that: The steps of counting the abnormal density of component states in the transmission line in each spatial region to obtain abnormal component hot zones, and raising the acceptance grade of the abnormal component hot zones are specifically: Count the number of abnormal status components in the transmission lines in each spatial area and the total number of components in the area; Calculating the ratio of the number of components with abnormal status to the total number of components in the region to obtain the abnormal density of components in the transmission line in each spatial region; Setting an abnormal density threshold, and comparing the abnormal density of the component with the abnormal density threshold; If the component abnormality density exceeds the abnormality density threshold, the spatial area is determined to be a component abnormality hot zone; The acceptance level of the abnormal hot zone of the component is raised, and the acquisition image quality threshold and the abnormality credibility threshold are increased.

9. A transmission line project acceptance system based on drones, characterized in that: The method for acceptance of a transmission line project based on a drone as described in any one of claims 1 to 8 comprises: The acceptance level setting module obtains satellite remote sensing data of the transmission line project, divides the transmission line into different spatial areas, and sets different acceptance levels for different spatial areas; Build an image acquisition module to control the drone according to the set flight path to collect images of various engineering components on the transmission line and obtain component image information; An abnormal credibility scoring module performs recognition analysis on the component image information to obtain a preliminary recognition result, and performs abnormal credibility scoring on the preliminary recognition result to obtain an abnormal credibility value; a reverse path triggering module, which triggers a reverse flight path signal based on the abnormal credibility value and controls the UAV to collect a reverse path image of the component; Constructing an abnormality confirmation module to perform recognition analysis on the reverse path image to obtain a reverse recognition result, judging the consistency between the preliminary recognition result and the reverse recognition result, and confirming that the component state is abnormal; The acceptance level adjustment module counts the abnormal density of component states in the transmission line in each spatial area, obtains component abnormal hot zones, and increases the acceptance level of the component abnormal hot zones.

Images