Abstract: For online damage detection for product packages, this paper proposes two methods based on coordinate transformation. We first select an image region to be analyzed, and obtain binary images using global threshold segmentation. Then, according to the image features, we rotate the coordinate system, and establish two damage recognition criteria, variance criterion and peak number criterion, based on the degree of separation and the probability distribution of projection values of the object pixels in the image. Finally, an image online detection system is set up. The package images are captured to verify the two criteria. The results show that both methods are effective and robust. The proposed technique can be used to similar online detection applications.

Key words: quality inspection, packaging detection, wrap carton detection, coordinate transformation, image processing