Inspection of Car Body Gaps Based on Stereo Vision
The car body gaps affect the vehicle sealing, in turn affect the performance of windproof and rainproof. Traditional inspection method cannot meet the requirement of on-line detection. The non-contact detection method is proposed for the car body gaps inspection in this paper. The gap area is taken by two cameras, and the information is obtained by image processing. The key is contour extraction, and the Snake method based on non-initialized level set is used in this article. The finally contour extraction result is exported by updating the profile curve. The experimental results show that the gaps inspection method based on stereo vision can meet the requirements of high-precision and not-contact detection, and the absolute errors of width and flush are 0.0605mm, 0.0712mm respectively. Furthermore, it can meet the requirements of on-line inspection.