Reserch On Key Technologies Of Image Preprocessing For Intelligent Patrol Inspecttion In Relay Protection Room Of Electricity Substation

Intelligent operation and maintenance is an important development trend of power inspection.The power inspection robot equipped with a visual acquisition terminal gradually replaces the manual inspection in the ancillary equipment inspection of the relay protection room.However,due to the complex inspection environment of the power inspection robot,higher requirements are submitted for the digital image processing algorithm.Aiming at the influence of complex imaging environment such as too small field of view,imaging observation angle offset,and high light target occlusion on imaging quality in complex imaging environment during patrol inspection,this paper focuses on preprocessing technologies such as image distortion correction,image mosaic,and image fusion,high light suppression.This paper solves the problems of visual imaging in complex inspection environments.It provides a better image source for subsequent image recognition and effectively improves the recognition rate.The main work of this paper is as follows:(1)Based on the image distortion correction and image mosaic algorithm,this paper proposes an image mosaic method based on feature block according to the characteristics of the substation relay room control cabinet.The experimental results show that the method used in this paper can effectively solve the problem of a large field of view imaging in a small space of substation relay room,and can realize the image distortion correction and large field of view imaging in a small space.Seamless stitching and image object clustering can obtain a clear image of the control panel.(2)The flexible adjustment of the visual imaging posture of the mobile robot is intended to achieve the best shooting of the target presentation through the feature extraction and matching of the target,the pre-estimation of the posture,and the flexible adjustment of the online control.First,obtain the camera sequence image information,and then perform target recognition processing on the read video,and label the target in the image frame to locate the target;finally,the target image position between frames is used as a constraint to adaptively adjust the robot’s Observe the attitude.(3)Based on the method of image segmentation and image fusion,the glass cabinet door highlights detection and suppression.First,detect whether the reference image has a highlight area,and then use the ORB algorithm to extract the feature point pairs of the reference image with the highlighted area and the image to be registered,and use the Hamming distance to eliminate the mismatched feature points to realize the rough removal of the feature points.Then the rough feature points are eliminated again based on the RANSAC algorithm to complete the accurate matching of the feature points;finally,the weighted fusion algorithm is used to effectively remove the highlight interference area of the platen image.The experimental results show that the method in this paper effectively solves the problems of complex imaging environment,such as too small field of view,imaging observation angle offset,and highlight target occlusion during the inspection process of the relay protection room of the substation,and improves the imaging quality of the robot vision terminal.The effect of inspection is of great significance to the promotion of intelligent inspection of substations.