Electric patrol inspection control method and system based on unmanned aerial vehicle

An electric patrol inspection control system based on an unmanned aerial vehicle, comprising:

• an unmanned aerial vehicle control module, configured to formulate a flight path according to an electric patrol inspection area, and to collect an image and position information of an electric device by a plurality of types of devices equipped on the unmanned aerial vehicle, wherein the image and the position information are transmitted to a receiving terminal of a ground station in a form of wireless transmission;
• an image processing and analyzing module, configured to receive the real-time collected image and position information by the ground station, to process and analyze the image, and then to obtain analysis on defects, damages and corrosion of the electric device;
• and
• a patrol and examine electronic reports generation module, configured to generate a patrol inspection electronic report comprising the analysis situations of a plurality of electric devices according to the analysis situation of the ground station and the position information as a distinguishing characteristic, to give the service life, maintenance expense, maintenance measure of the electric device at the same time according to the patrol inspection electronic report, and to inform maintenance personnel of the position information of the electric device;
• wherein the unmanned aerial vehicle control module comprises:
• a position information acquisition submodule, configured to acquire an electric device distribution map in the electric patrol inspection area, and at the same time to acquire position coordinates of various electric devices, wherein the position coordinates are the position information, the plurality of electric devices are divided into a plurality of flight path nodes according to the different position information, and the flight path nodes comprising the position information constitute a training trajectory dataset;
• a flight path generation submodule, configured to input the training trajectory dataset to a flight path generation model, to generate a flight path in real time, and to input the flight path to a controller of the unmanned aerial vehicle;
• a patrol inspection time minimization submodule, configured to minimize time that the unmanned aerial vehicle completes the specified electric patrol inspection area in combination with wireless transmission power and the flight path of the unmanned aerial vehicle, wherein a trial flight is performed on the unmanned aerial vehicle according to the flight path, to check the time;
• a first background field formation submodule, configured to continue the plurality of flight path nodes to obtain a flight route of the adjacent flight path nodes, to divide the flight route into a plurality of grids, and to construct a flight path background field comprising the plurality of grids;
• a second background field formation submodule, configured to acquire geographical environment data of the electric patrol inspection area by the laser radar, to preprocess the geographical environment data, to construct a geographical environment background field of the electric patrol inspection area, and to divide the geographical environment background field into a plurality of grids;
• and
• a fly-around judgment submodule, configured to merge the grids of the flight path background field with the grids of the geographical environment background field to form a grid matrix, to judge a significance of a flight direction of the unmanned aerial vehicle based on the grids in the grid matrix, and to confirm whether fly-around is required or not in the current geographical environment;
• wherein the image processing and analysis module comprises:
• an image preprocessing submodule, configured to preprocess the received image, wherein the preprocessing comprises operations of image denoising, enhancing a contrast ratio and adjusting a brightness, the electric device in the image is subjected to object detection and segmentation, to identify and position a position and a bounding box of the electric device;
• a characteristic vector acquisition submodule, configured to extract a target characteristic related to a defect, damage and corrosion in a segmented electric device area, and to transform the image information into a characteristic vector capable of being quantified and analyzed, wherein the target characteristic comprises a texture characteristic, a shape characteristic and a color characteristic;
• a situation judgment submodule, configured to perform a defect analysis and a classification on the extracted target characteristic by using a classifier, to judge the defect, damage and corrosion situations of the electric device according to the characteristic vector, and to give a corresponding judgment result;
• wherein extracting the target characteristic related to the defect, damage and corrosion from the characteristic vector acquisition submodule comprises:
• identifying the target characteristic of the characteristic in the segmented electric device area, extracting the target characteristic of a segmented to-be-identified electric device image and a segmented preset electric device image, to obtain a plurality of characteristic vectors, wherein the characteristic vectors correspond to texture characteristic, shape characteristic and color characteristic information of the segmented to-be-detected electric device image;
• calculating a similarity of the characteristic vectors of the segmented to-be-identified electric device image and the segmented preset electric device image under the texture characteristic, shape characteristic and color characteristic of the same electric device image, to obtain a first similarity of the texture similarity, a second similarity of the shape similarity and a third similarity of the color similarity, wherein a computed expression of the similarity Sim(Lp,Lq) is
• Sim
• ⁢
• (
• L
• p
• ,
• L
• q
• ,
• L
• s
• )
• =
• 2
• ⁢
• ∑
• k
• =
• 1
• N
• ⁢
• min
• ⁡
• (
• i
• pk
• ,
• i
• qk
• ,
• i
• s
• ⁢
• k
• )
• ∑
• k
• =
• 1
• N
• ⁢
• (
• i
• p
• ⁢
• k
• +
• i
• q
• ⁢
• k
• +
• i
• s
• ⁢
• k
• )
• ;
• wherein Lp, Lq, Ls correspond to the characteristic vectors of the texture characteristic, shape characteristic and color characteristic in respective, ipk represents a kth element of the texture characteristic vector Lp, iqk represents a kth element of the shape characteristic vector Lp, isk represents a kth element of the color characteristic vector Ls, min(⋅) represents a function taking the minimum, N represents the number of elements in the characteristic vectors, specifically for the texture characteristic, shape characteristic and color characteristic, each element in the characteristic vectors corresponds to a certain specific quantified quantity or statistic quantity of the respective characteristic;
• for the texture characteristic, each element possibly represents a value of a certain texture statistic indicator;
• for the shape characteristic, each element possibly represents a value of a certain geometrical characteristic;
• and for the color characteristic, each element possibly represents a value of a certain color attribute;
• and
• obtaining a difference of the electric device image in the texture characteristic, shape characteristic and color characteristic according to the first similarity, the second similarity and the third similarity that are obtained upon calculation, thus obtaining the change of the target characteristic related to the defect, damage and corrosion.