Unmanned aerial vehicle-based distribution network non-power-cut operation digital investigation method

Abstract

The invention discloses an unmanned aerial vehicle-based distribution network uninterruptible power operation digital investigation method, which comprises the following steps of: dividing an investigation area into a plurality of flight areas with the same area, monitoring and collecting the flight path of an unmanned aerial vehicle, performing investigation evaluation according to the flight path information, preliminarily identifying the hidden danger of insufficient accuracy of image data, and obtaining the unmanned aerial vehicle-based distribution network uninterruptible power operation digital investigation result. Then feature extraction and analysis are carried out on each flight area, an image overlapping index and an image quality index are calculated, an area image quality coefficient is generated through a machine learning model, the flight area is divided into a high-quality investigation area and a low-quality investigation area based on the coefficient, and a flight strategy of the low-quality area is optimized; according to the invention, through scientific division of the investigation area, real-time monitoring of the flight path, evaluation of the data quality, generation of the image quality coefficient by using the machine learning model, and optimization of the flight strategy, efficient and safe distribution network non-power-cut operation digital investigation can be realized, and the investigation data has enough accuracy and reliability.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) applications, and in particular to a digital survey method for uninterrupted power supply operations in a distribution network based on UAVs. Background Art

[0002] Uninterrupted operation is an important means to improve power supply reliability and service quality. However, the current manual survey method has problems such as low efficiency and incomplete information collection. As an emerging technical means, drones have been widely used in various fields due to their flexibility and efficiency, especially in the inspection and survey of power systems. The unmanned aerial vehicle-based distribution network uninterrupted operation digital survey method can significantly improve the survey efficiency and data quality, reduce labor costs and safety risks, and reduce the cost of manpower and safety risks by introducing drone technology. However, during the drone survey process, the accuracy and quality of the data directly affect the survey effect.

[0003] At present, the existing digital survey methods for distribution network non-stop operations based on drones all have problems with the accuracy and reliability of drone survey data, and cannot guarantee that their survey data have sufficient accuracy and reliability. This has also become an important direction of current research. Summary of the invention

[0004] The purpose of the present invention is to provide a digital survey method for distribution network non-stop operation based on drones to solve the above defects.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] A digital survey method for distribution network non-stop operation based on drone, comprising the following steps:

[0007] S1. Divide the survey area into multiple flight areas of equal size, monitor and collect the flight trajectory of the drone, and conduct a survey and assessment of the survey area based on the monitored flight trajectory information to preliminarily identify whether the survey area has hidden dangers of insufficient image data accuracy;

[0008] S2. Based on the results of the survey and evaluation, the survey features are extracted for each flight area, and then feature analysis is performed to obtain an image overlap index and an image quality index, and the image overlap index and the image quality index are transmitted to a pre-trained machine learning model, and a regional image quality coefficient is generated through the pre-trained machine learning model;

[0009] S3, based on the generated regional image quality coefficient, dividing each flight area into a high-quality survey area and a low-quality survey area;

[0010] S4. Optimize the flight strategy of UAV surveys in low-quality survey areas.

[0011] Preferably, in step S1, a survey and evaluation is performed on the survey area to preliminarily identify whether the survey area has a hidden danger of insufficient image data accuracy. The specific operations are as follows:

[0012] According to the monitored flight trajectory information, obtain the timestamp of the entry and exit time of the drone in each flight area, get the time difference of each entry and exit as the stay time value of each time, and then sum up all the stay time values ​​in the target flight area to get the survey time value of the target flight area;

[0013] The survey average and survey standard deviation are obtained through all the survey time values, and the survey average is compared with the preset standard average, and the survey standard deviation is compared with the preset standard deviation benchmark value. If the survey average is greater than or equal to the preset standard average and the survey standard deviation is less than the preset standard deviation benchmark value, a normal signal is generated. If the survey average is not greater than or equal to the preset standard average and the survey standard deviation is less than the preset standard deviation benchmark value, an abnormal signal is generated. When an abnormal signal is generated, it is preliminarily identified whether the survey area has the hidden danger of insufficient image data accuracy.

[0014] Preferably, in step S2, survey feature extraction is performed on each flight area, so that the features obtained are divided into image overlap information features and image acquisition quality information features.

[0015] Preferably, in step S3, the image overlap index is obtained by the following logic:

[0016] Get the image sequence X collected in the target flight area: I1, I2, ..., I n ; n is the total number of images collected, the width of each image is W, the height is H, for each image I i Use SIFT algorithm to detect feature points and match adjacent images I i and I j feature points to obtain matching pairs j=i+1, Represents image I i The kth feature point detected on Represents image I j The lth feature point detected on ;

[0017] Then calculate the overlap area Q and the number of pixels in the overlap area NX, and then calculate the image I i and I j The overlap rate R between them is calculated as: NZ is the total number of pixels of a single image, satisfying the following condition: NZ = W*H. The overlap ratio R of all adjacent graphics is averaged to obtain the image overlap index.

[0018] Preferably, the acquisition logic of the overlapping area Q is:

[0019] The RANSAC algorithm is used to estimate the homography matrix T between adjacent images and perform registration. The homography matrix T is used to convert image I j Register to image I i Coordinate system; use RANSAC algorithm based on matching feature point pairs Estimate the homography matrix T, describing the image I j To Image I i Mapping, using the homography matrix T to the image I j Perform perspective transformation and align it to image I i In the coordinate system of , the image I is determined by perspective transformation. j The four corner points of image I i The position in the image I is calculated by calculating the transformed corner points and image I i The intersection of the boundaries results in the overlapping area Q.

[0020] Preferably, in step S2, the image quality index is obtained by the following logic:

[0021] Get the image sequence X collected in the target flight area: I1, I2, ..., I n ; n is the total number of images collected, the width of each image is W, the height is H, and then the spatial frequency value is calculated using the following formula:

[0022]

[0023] In the formula, SF x and SF y Respectively represent the spatial frequency of the image in the horizontal and vertical directions, SF is the spatial frequency value;

[0024] Calculate the natural scene value, the calculation formula is:

[0025]

[0026] In the formula, i is the index of each pixel in the image, N is the total number of pixels in the image, and I i is the gray value of the i-th pixel, p(I i ) is the pixel gray value I i The probability distribution of , NSS is the natural scene value;

[0027] Calculate the information entropy value, the calculation formula is:

[0028]

[0029] In the formula, i is the gray level index, L is the number of gray levels, and p i is the probability of gray level i in the image, and H is the information entropy value;

[0030] The image quality features: spatial frequency value, natural scene value, and information entropy value are normalized respectively, and then the geometric mean method is used to integrate the normalized image quality features to calculate the image quality index.

[0031] Preferably, the SF x , SF y The acquisition logic is as follows:

[0032] SF x The calculation formula is:

[0033]

[0034] SF y The calculation formula is:

[0035]

[0036] In the formula, is the horizontal gradient of the image, is the vertical gradient of the image, i and j represent the row and column indices of the pixel in the image respectively.

[0037] Preferably, the step S3 is specifically as follows: obtaining a regional image quality coefficient of the target flight area, comparing it with a preset survey quality threshold, and if the regional image quality coefficient of the target flight area is greater than or equal to the preset survey quality threshold, classifying the target flight area as a high-quality survey area; if the regional image quality coefficient of the target flight area is less than the preset survey quality threshold, classifying the target flight area as a low-quality survey area.

[0038] Preferably, the step S4 is specifically as follows: for low-quality survey areas, the A algorithm is used to optimize the UAV survey path, and the survey image is reacquired in combination with an adaptive aircraft attitude adjustment strategy.

[0039] The beneficial effects of the present invention are:

[0040] (1) The digital survey method for non-stop power supply operation of distribution network based on UAV of the present invention can realize efficient, accurate and safe digital survey for non-stop power supply operation of distribution network by scientifically dividing the survey area, monitoring the flight trajectory in real time, evaluating the data quality, generating the image quality coefficient by using the machine learning model, and optimizing the flight strategy.

[0041] (2) The present invention is based on a digital survey method for distribution network without power outages using drones. By real-time monitoring and evaluation of the survey area, the drone's residence time in each flight area is analyzed, and the average and standard deviation of the survey time are calculated to ensure the accuracy and consistency of the survey data. The potential danger areas where the image data is not accurate enough are preliminarily identified, and timely adjustments and optimizations are made to improve the reliability of the data. The potential dangers of insufficient image data accuracy can be identified in a timely manner to ensure data accuracy. The flight strategy is optimized for low-quality survey areas to ensure that the drone performs sufficient data collection in these areas, thereby improving the overall survey effect. A scientific data quality evaluation system is provided through analysis of the image overlap index and the image quality index combined with the regional image quality coefficient generated by the machine learning model.

[0042] (3) The present invention provides a digital survey method for distribution network without power outages based on drones. Through an automated flight trajectory monitoring and evaluation system, it reduces reliance on manual intervention, improves the efficiency and safety of survey work, and can ensure that surveys are carried out without power outages in the distribution network, thereby reducing the economic losses and inconveniences caused by power outages and improving the reliability of power grid operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 : The principle diagram of the digital survey method for distribution network non-stop operation based on drone of the present invention. DETAILED DESCRIPTION

[0044] The present invention is further described below in conjunction with the embodiments. It should be noted that these are merely examples and illustrations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the claims, they should be deemed to fall within the protection scope of the present invention.

[0045] Embodiment 1:

[0046] Figure 1 The schematic diagram of the digital survey method for non-stop power supply operation of distribution network based on UAV of the present invention is as follows: Figure 1 As shown, a distribution network non-stop operation digital survey method based on drones includes the following steps:

[0047] S1. Divide the survey area into multiple flight areas of equal size, monitor and collect the flight trajectory of the drone, and conduct a survey and assessment of the survey area based on the monitored flight trajectory information to preliminarily identify whether the survey area has hidden dangers of insufficient image data accuracy;

[0048] S2. Based on the results of the survey and evaluation, the survey features are extracted for each flight area, and then feature analysis is performed to obtain an image overlap index and an image quality index, and the image overlap index and the image quality index are transmitted to a pre-trained machine learning model, and a regional image quality coefficient is generated through the pre-trained machine learning model;

[0049] S3, based on the generated regional image quality coefficient, dividing each flight area into a high-quality survey area and a low-quality survey area;

[0050] S4. Optimize the flight strategy of UAV surveys in low-quality survey areas.

[0051] Embodiment 2:

[0052] A method for digital survey of distribution network without power outage based on drone, the steps of which are the same as those in Example 1, and further include:

[0053] In step S1, the survey area is surveyed and evaluated to preliminarily identify whether the survey area has hidden dangers of insufficient image data accuracy. The specific operations are as follows:

[0054] According to the monitored flight trajectory information, obtain the timestamp of the entry and exit time of the drone in each flight area, get the time difference of each entry and exit as the stay time value of each time, and then sum up all the stay time values ​​in the target flight area to get the survey time value of the target flight area;

[0055] The survey average and survey standard deviation are obtained through all the survey time values, and the survey average is compared with the preset standard average, and the survey standard deviation is compared with the preset standard deviation benchmark value. If the survey average is greater than or equal to the preset standard average and the survey standard deviation is less than the preset standard deviation benchmark value, a normal signal is generated. If the survey average is not greater than or equal to the preset standard average and the survey standard deviation is less than the preset standard deviation benchmark value, an abnormal signal is generated. When an abnormal signal is generated, it is preliminarily identified whether the survey area has the hidden danger of insufficient image data accuracy.

[0056] Embodiment 3:

[0057] A method for digital survey of distribution network without power outage based on drone, the steps of which are the same as those of embodiment 1 or embodiment 2, and further include:

[0058] In step S2, survey feature extraction is performed on each flight area, and the features obtained are divided into image overlap information features and image acquisition quality information features.

[0059] Among them, the image quality index, its acquisition logic is as follows:

[0060] Get the image sequence X collected in the target flight area: I1, I2, ..., I n ; n is the total number of images collected, the width of each image is W, the height is H, and then the spatial frequency value is calculated using the following formula:

[0061]

[0062] Where SF x and SF y Respectively represent the spatial frequency of the image in the horizontal and vertical directions, SF is the spatial frequency value;

[0063] Calculate the natural scene value, the calculation formula is:

[0064]

[0065] In the formula, i is the index of each pixel in the image, N is the total number of pixels in the image, and I i is the gray value of the i-th pixel, p(I i ) is the pixel gray value I i The probability distribution of , NSS is the natural scene value;

[0066] Calculate the information entropy value, the calculation formula is:

[0067]

[0068] In the formula, i is the gray level index, L is the number of gray levels, and p i is the probability of gray level i in the image, and H is the information entropy value;

[0069] The image quality features: spatial frequency value, natural scene value, and information entropy value are normalized respectively, and then the geometric mean method is used to integrate the normalized image quality features to calculate the image quality index.

[0070] Among them, SF x , SF y The acquisition logic is as follows:

[0071] SF x The calculation formula is:

[0072]

[0073] SF y The calculation formula is:

[0074]

[0075] In the formula, is the horizontal gradient of the image, is the vertical gradient of the image, i and j represent the row and column indices of the pixel in the image respectively.

[0076] Embodiment 4:

[0077] A method for digital survey of distribution network without power outage based on drone, the steps of which are the same as those of embodiment 1, embodiment 2 or embodiment 3, and further include:

[0078] The step S3 is specifically as follows: obtaining a regional image quality coefficient of the target flight area, comparing it with a preset survey quality threshold, and if the regional image quality coefficient of the target flight area is greater than or equal to the preset survey quality threshold, then classifying the target flight area as a high-quality survey area; if the regional image quality coefficient of the target flight area is less than the preset survey quality threshold, then classifying the target flight area as a low-quality survey area.

[0079] In step S3, the image overlap index is obtained by the following logic:

[0080] Get the image sequence X collected in the target flight area: I1, I2, ..., I n ; n is the total number of images collected, the width of each image is W, the height is H, for each image I i Use SIFT algorithm to detect feature points and match adjacent images I i and I j feature points to obtain matching pairs j=i+1, Represents image I i The kth feature point detected on Represents image I j The lth feature point detected on ;

[0081] Then calculate the overlap area Q and the number of pixels in the overlap area NX, and then calculate the image I i and I j The overlap rate R between them is calculated as: NZ is the total number of pixels of a single image, satisfying the following condition: NZ = W*H. The overlap ratio R of all adjacent graphics is averaged to obtain the image overlap index.

[0082] Among them, the overlapping area Q, its acquisition logic is:

[0083] The RANSAC algorithm is used to estimate the homography matrix T between adjacent images and perform registration. The homography matrix T is used to convert image I j Register to image I i Coordinate system; use RANSAC algorithm based on matching feature point pairs Estimate the homography matrix T, describing the image Ij To Image I i Mapping, using the homography matrix T to the image I j Perform perspective transformation and align it to image I i In the coordinate system of , the image I is determined by perspective transformation. j The four corner points of image I i The position in the image I is calculated by calculating the transformed corner points and image I i The intersection of the boundaries results in the overlapping area Q.

[0084] Embodiment 5:

[0085] A method for digital survey of distribution network without power outage based on drone, the steps of which are the same as those of embodiment 1, embodiment 2, embodiment 3 or embodiment 4, and further include:

[0086] Step S4 is as follows: For low-quality survey areas, the A algorithm is used to optimize the UAV survey path, and the survey image is reacquired in combination with an adaptive aircraft attitude adjustment strategy.

[0087] The present invention is based on the digital survey method for distribution network non-stop operation of unmanned aerial vehicles. By scientifically dividing the survey area, real-time monitoring of the flight trajectory, evaluating the data quality, using the machine learning model to generate the image quality coefficient, and optimizing the flight strategy, efficient, accurate and safe digital survey for distribution network non-stop operation can be achieved.

[0088] The present invention is based on a digital survey method for unmanned aerial vehicle (UAV) distribution network without power outages. The method performs real-time monitoring and evaluation on the survey area, analyzes the residence time of the UAV in each flight area, calculates the average and standard deviation of the survey time, ensures the accuracy and consistency of the survey data, preliminarily identifies potential danger areas with insufficient image data accuracy, makes timely adjustments and optimizations, improves data reliability, can timely identify potential dangers with insufficient image data accuracy, ensures data accuracy, optimizes flight strategies for low-quality survey areas, ensures that the UAV performs sufficient data collection in these areas, improves the overall survey effect, and provides a scientific data quality evaluation system through analysis of the image overlap index and the image quality index combined with the regional image quality coefficient generated by a machine learning model.

[0089] The present invention is based on the digital survey method for distribution network non-stop operation of unmanned aerial vehicles. Through the automated flight trajectory monitoring and evaluation system, the reliance on manual intervention is reduced, the efficiency and safety of the survey work are improved, and it can ensure that the survey is carried out without power outages in the distribution network, thereby reducing the economic losses and inconveniences caused by power outages and improving the reliability of power grid operation and maintenance.

[0090] The above is an exemplary description of the invention. Obviously, the specific implementation of the present invention is not limited to the above-mentioned method. As long as such non-substantial improvements are made by adopting the method concept and technical solution of the present invention, or the concept and technical solution of the invention are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.

Claims

1. A digital survey method for distribution network non-stop operation based on drone, characterized in that: The following steps are involved: S1. Divide the survey area into multiple flight areas of equal size, monitor and collect the flight trajectory of the drone, and conduct a survey and assessment of the survey area based on the monitored flight trajectory information to preliminarily identify whether the survey area has hidden dangers of insufficient image data accuracy; S2. Based on the results of the survey and evaluation, the survey features are extracted for each flight area, and then feature analysis is performed to obtain an image overlap index and an image quality index, and the image overlap index and the image quality index are transmitted to a pre-trained machine learning model, and a regional image quality coefficient is generated through the pre-trained machine learning model; S3, based on the generated regional image quality coefficient, dividing each flight area into a high-quality survey area and a low-quality survey area; S4. Optimize the flight strategy of UAV surveys in low-quality survey areas.

2. The method for digital survey of distribution network without power outage based on drone according to claim 1 is characterized in that: In step S1, the survey area is surveyed and evaluated to preliminarily identify whether the survey area has a hidden danger of insufficient image data accuracy. The specific operations are as follows: According to the monitored flight trajectory information, obtain the timestamp of the entry and exit time of the drone in each flight area, get the time difference of each entry and exit as the stay time value of each time, and then sum up all the stay time values ​​in the target flight area to get the survey time value of the target flight area; The survey average and survey standard deviation are obtained through all the survey time values, and the survey average is compared with the preset standard average, and the survey standard deviation is compared with the preset standard deviation benchmark value. If the survey average is greater than or equal to the preset standard average and the survey standard deviation is less than the preset standard deviation benchmark value, a normal signal is generated. If the survey average is not greater than or equal to the preset standard average and the survey standard deviation is less than the preset standard deviation benchmark value, an abnormal signal is generated. When an abnormal signal is generated, it is preliminarily identified whether the survey area has the hidden danger of insufficient image data accuracy.

3. The method for digital survey of distribution network without power outage based on drone according to claim 1 is characterized in that: In step S2, survey feature extraction is performed on each flight area, and the features obtained are divided into image overlap information features and image acquisition quality information features.

4. The method for digital survey of distribution network without power outage based on drone according to claim 1 is characterized in that: In step S3, the image overlap index is obtained by the following logic: Get the image sequence X collected in the target flight area: I1, I2, ..., I n ; n is the total number of images collected, the width of each image is W, the height is H, for each image I i Use SIFT algorithm to detect feature points and match adjacent images I i and I j feature points to obtain matching pairs j=i+1, Represents image I i The kth feature point detected on Represents image I j The lth feature point detected on ; Then calculate the overlap area Q and the number of pixels in the overlap area NX, and then calculate the image I i and I j The overlap rate R between them is calculated as: NZ is the total number of pixels of a single image, satisfying the following condition: NZ = W*H. The overlap ratio R of all adjacent graphics is averaged to obtain the image overlap index.

5. The method for digital survey of distribution network without power outage based on drone according to claim 4 is characterized in that: The acquisition logic of the overlapping area Q is: The RANSAC algorithm is used to estimate the homography matrix T between adjacent images and perform registration. The homography matrix T is used to convert image I j Register to image I i The coordinate system of Use the RANSAC algorithm based on matching feature point pairs Estimate the homography matrix T, describing the image I j To Image I i Mapping, using the homography matrix T to the image I j Perform perspective transformation and align it to image I i In the coordinate system of , the image I is determined by perspective transformation. j The four corner points of image I i The position in the image I is calculated by calculating the transformed corner points and image I i The intersection of the boundaries results in the overlapping area Q.

6. The method for digital survey of distribution network without power outage based on drone according to claim 1 is characterized in that: In step S2, the image quality index is obtained by the following logic: Get the image sequence X collected in the target flight area: I1, I2, ..., I n ; n is the total number of images collected, the width of each image is W, the height is H, and then the spatial frequency value is calculated using the following formula: Where SF x and SF y Respectively represent the spatial frequency of the image in the horizontal and vertical directions, SF is the spatial frequency value; Calculate the natural scene value, the calculation formula is: Where i is the index of each pixel in the image, N is the total number of pixels in the image, and I i is the gray value of the i-th pixel, p(I i ) is the pixel gray value I i The probability distribution of , NSS is the natural scene value; Calculate the information entropy value, the calculation formula is: In the formula, i is the gray level index, L is the number of gray levels, and p i is the probability of gray level i in the image, and H is the information entropy value; The image quality features: spatial frequency value, natural scene value, and information entropy value are normalized respectively, and then the geometric mean method is used to integrate the normalized image quality features to calculate the image quality index.

7. The method for digital survey of distribution network without power outage based on drone according to claim 6 is characterized in that: The SF x , SF y The acquisition logic is as follows: SF x The calculation formula is: SF y The calculation formula is: In the formula, is the horizontal gradient of the image, is the vertical gradient of the image, i and j represent the row and column indices of the pixel in the image respectively.

8. The method for digital survey of distribution network without power outage based on drone according to claim 1 is characterized in that: The S3 step is specifically as follows: The regional image quality coefficient of the target flight area is obtained and compared with the preset survey quality threshold. If the regional image quality coefficient of the target flight area is greater than or equal to the preset survey quality threshold, the target flight area is divided into a high-quality survey area; if the regional image quality coefficient of the target flight area is less than the preset survey quality threshold, the target flight area is divided into a low-quality survey area.

9. The method for digital survey of distribution network without power outage based on drone according to claim 1, characterized in that: The S4 step is specifically as follows: For low-quality survey areas, the A algorithm is used to optimize the UAV survey path, and the adaptive aircraft attitude adjustment strategy is combined to reacquire the survey image.

Images