Application of Fast Parallel Thinning Algorithm in Seam Tracking

Since distinguishing features are not obvious between the welding seam quality testing section and other parts, there is a lot of noise in the image acquisition which leads deviations for weld identification. The authors put forward a new method to identify the appropriate solutions. This paper presents the approaches for image processing steps such as images obtaining, image gray transformation, image binarization, image reduction, and image fast parallel thinning. The programming language C# is used to implement weld seam for effective image recognition. The experimental results show that the method can achieve fast recognition, attain high accuracy and is easy to transplant.