Research On The Image Processing Technology For Automobile Headlamp Testing | | Posted on:2008-04-30 | Degree:Master | Type:Thesis | | Country:China | Candidate:H Y Zhu | Full Text:PDF | | GTID:2132360215976495 | Subject:Carrier Engineering | | Abstract/Summary: | PDF Full Text Request | | The automobile headlamp is one of the important parts for safe driving. The headlamp beam direction testing is the legal item of the vehicle safety performance testing in current national standard. The existing methods and equipments have the problems of low precision, poor repeatability and high misjudgment rate, which can not meet the need of vehicle safety performance testing. In order to improve the testing method of headlamp beam direction with two measuring screens and the correction method of automobile location error based on image processing technology, the theory and methods of automatic image acquiring and processing are researched by theoretical analysis and experiment in this paper.A fast median filtering algorithm is used to preprocess the headlamp low beam projection image. The median filter processing speed is increased. According to the characteristic of gray histogram for headlamp low beam projection image, an improved Otsu segmentation method is adopted to improve the quality and speed of automatic segmentation. The least square method is adopted for straight line fitting to fix position of turning point of the boundary between bright part and dim part of low beam projection. The testing speed and precision of low beam azimuth are enhanced.According to the characteristic of approach method for figuring out the automobile directaxis, a two-dimensional maximum entropy threshold value segmentation algorithm based on image partitioning is proposed, which can improve the adaptability for automobile image with uneven illumination, and increase the image processing speed.The software of image acquiring and processing is developed with MATLAB and Visual C++. The system can realize automatic testing for low beam azimuth and self-correcting for automobile location error.The experiment system for the low beam azimuth testing and automobile location error correction is developed and experimentized. The experiment results show that the testing method of headlamp beam direction with two measuring screens and the correction method of automobile location error based on image processing technology are feasible, and the practicability of above two methods is improved by the software. | | Keywords/Search Tags: | headlamp testing, image acquiring, image segmentation, Otsu, 2-D maximum entropy threshold method, MATLAB, Visual C++ | PDF Full Text Request | Related items |
| |
|