Recognition And Location Of Automatic Charging Interface Of Electric Velhicle Based On Image Recognition Algorithm And Its Control Method

With the development of science and technology and people increasingly demanding environmental protection,electric vehicles increasingly popular.But in the manual electric vehicle charging process,the operation is more cumbersome,there may be equipment aging or damage caused by security risks,so improve the degree of automation of electric vehicle charging process is very necessary.Based on the research of mechanical structure control and image processing technology,this paper presents a set of automatic charging system for electric vehicle,including the hardware structure,software system and the realization of docking core image recognition algorithm.In this paper,according to the electric car charging process of the automatic charging system hardware structure structures.Including vehicle and body posture detection device,charging target identification and docking device,automatic charging control device,charging state monitoring device.Through a long range of three-dimensional mobile platform,making the automatic charging system has a high applicability and flexibility,and the structure is simple,easy to control,low cost,easy to damage.In this paper,according to the process of automatic control of electric vehicles in the control process and functional requirements for the preparation of the software system.Including vehicle detection function,body posture detection function,charging target recognition and positioning function,charging end docking function,charging state monitoring function.In this paper,the dynamic proportional parameter PD control method can effectively speed up the stabilization process and reduce the steady-state error.The use of monocular vision sensor combined with closed-loop control principle through the calibration-movement cycle to form closed-loop feedback control,to obtain a higher positioning accuracy.This paper focuses on the image processing algorithm to realize the target recognition and location of charging interface.The whole image processing process includes image preprocessing and target recognition and location.According to the different characteristics of different images,the target recognition and positioning process is divided into two stages,and different recognition and positioning algorithms are selected for different characteristics of different stages.Increases the recognition rate and positioning accuracy for each stage.In the preliminary docking stage,the Hough circle transform is used to carry out the contour extraction according to the characteristics of the gradient feature of the image and the obvious contour feature.After the canny edge detection,the contour recognition is carried out,and a practical application scenario is proposed.Adaptive Hough circle transformation method and false target elimination method based on confidence.Compared with the traditional Hough transform,this method improves the error caused by improper parameter setting by adaptive parameter calculation and confidence screening.Sentenced to a higher recognition rate.In the accurate docking stage with close distance,the target location is proposed by matching the SURF feature points according to the characteristics of the image gradient.This paper is based on the design of the test system to build and experiment.For hardware installation and software deployment to complete the experimental system to build,and functional and performance testing.Based on the experimental scheme,the adaptive Hough circle transformation method is designed to achieve 96%recognition rate in the application of circular charge target contour recognition.And through the precise positioning method can achieve 3 pixel error within the target positioning,and complete the accurate charging interface plug.