Complex Background Image Detection and Processing Based on Machine Vision

This paper suggests a new method that generates sampling point along the guide line for linear contour defect detection and recognition. The method registers the vector data with the measured image to generate sampling points that can cover the contour accurately. Next, these sampling points are saved into a test file for quick batch detection of the same products. Then pictures are taken to eliminate the complex background noise on the objects to carry out a series of image pretreatment. On this basis, the algorithm that using SUSAN Operator is used to determine the location of the sampling points. Also, the potential defects are identified and classified. The method has good adaptability to the detection of linear image contour. It is of high application value due to reliability and real-time capability.