The Image Location System Of Auto Electric Seat’ Switch

Nowadays,automobile electric seats are detected by workers before leaving the factory and entering the market in China.Meanwhile,manually manipulating seats to a proper position for the following process operations when the detection is over.Artificial detection brings some questions,such as high intensity labor,low detection efficiency and bad stability.Therefore,machine vision and image processing are developed into automobile electric seats system and the core of it is image location system.Methods and contens of the study include the following aspects:(1)Image denoising.The environment of seats detection is complex,so captured seat images often bring kinds of noises to reduce the quality.Now,image denoising algorithms have become one of hot research directions.This paper has an in-depth study about three main algorithms based on wavelet theory,partial differential equation(PDE)and non-local means(NLM)algorithm.A kind of improved NLM algorithm combined bilateral filtering(BF)with discrete cosine transform(DCT)denoising method was proposed in this paper.Space adjacent function of BF is coupled with original NLM weight function to protect image detail information.And focused power character of DCT is used to calculate the pixel similarity for improving calculation speed.Experiment results show that compared with NLM algorithm,the proposed method protects image edge and detail information.(2)Image recognition and positioning.According to the work environment of seat production line,monocular vision detection system is developed to achieve the feature extraction of seat switches.Feature extraction includes edge extraction and shape extraction.Finally,using Hough Circle Transform completes the precise recognition.Meanwhile,Studying the relationship between space 3D objects and corresponding points in 2D image.And researching the camera imaging model and using the calibration algorithm to obtain camera inner and outer parameters.The location in imaging coordinate that instrument equipment obtains is transformed into world coordinate finally.(3)Image positioning system.The image positioning system of automobile electric seats is designed and developed.The system has a good graphical human-computer interaction interface integrating many functions: machine vision,image processing and database management.It can provide a precise object location for the manipulator.Also,the manipulator replaces workers’ hands to complete the detection task,which reducing labor intensity and improving detection efficiency mostly.