A Study on Image-Based On-line Detection of Folding Errors

Images of folding process was collected and on-line detection of folding dislocations is explored using characteristic points in these images, which reduces the computation time. First, determine the dislocation type through characteristic points in the images of a folding process. Then, through calibration the corresponding dislocation bias for the detected type of folding dislocation can be obtained. A simple experimental device was assembled. A CMOS camera was selected for image capture, and the parameters of the CMOS camera were set according to the characteristics of the experimental objects. The detection range for folding dislocations was determined taking into account requirements such as processing speeds, image sizes, and image resolutions. Experimental results show that a satisfactory on-line detection of folding dislocations is achieved.