Research On Recognition Algorithm And Exposure Algorithm Of Electric Vehicle Charging Socket

In recent years,the global electric vehicle ownership has grown very rapidly.How to charge efficiently and safely has always been a concern in the field of electric vehicles.But there are many disadvantages in manual charging,such as large work intensity,poor working environment,equipment with safety hazards,huge number of workers,and more and more difficult to recruit workers.These disadvantages have become bottlenecks restricting the efficient,stable,safe and benign operation of electric vehicles and public transport systems.Compared with manual operation,robot automatic charging can solve several problems of manual operation.In addition,with the popularization of automatic driving and automatic parking,automatic charging using robots will become an inevitable trend.Aiming at the difficulty of identifying multi-attitude charging sockets with simple structure and few features,the algorithm of recognizing charging sockets is studied.Meanwhile,in order to improve the robustness of the recognition algorithm to different illumination conditions,the charging socket exposure algorithm is studied.The recognition characteristics of the charging socket is analyzed.A charging socket recognition algorithm that combines the Hu invariant moment and the LBP texture feature to identify the charging socket is proposed.Firstly,the image preprocessing algorithm is designed to realize the acquisition of candidate regions that may contain charging sockets.Then,the charging socket recognition algorithm based on Hu invariant moment is designed to realize the preliminary screening of candidate regions.Finally,a charging socket recognition algorithm based on LBP texture features is designed to further screen the candidate areas,thus completing the recognition of charging sockets.The structure of the charging socket is analyzed.And a charging socket exposure algorithm based on Laplacian pyramid image fusion is proposed.Firstly,the exposure algorithm based on the image gray histogram is designed to obtain the clear image of the contact part of the charging socket.Then,the exposure algorithm based on the image gray level is designed to obtain the clear image of the outer contour of the charging socket.The image registration algorithm is designed to realize the image registration.Furthermore,the image fusion algorithm based on the Laplacian pyramid is designed to realize the fusion of the two exposed images after registration.C++ programming language is used to realize the software of recognition algorithm and exposure algorithm.Then the recognition algorithm and exposure algorithm are tested by experiments.Firstly,the recognition effect of recognition algorithm on different attitude charging sockets and the recognition accuracy of recognition algorithm are tested.Then,the quality of images acquired by the exposure algorithm in different scenes is evaluated by selecting evaluation criteria.The experimental results show that the charging sockets of various attitudes can be recognized within the required angle and distance range,and the recognition accuracy is very high.Fused images can combine the advantages of two images,and the entropy of the image increases by 3.91 % and 76.40 % respectively.The fusion image can completely present the detailed features of the charging socket,which is of great significance to improve the reliability and robustness of the recognition algorithm.