A Detecting Approach for Misplaced Packages during Palletization Based on Machine Vision

In this study, an inspection system is designed to detect the emergence of tilted placed titanium dioxide packages during palletization. An industrial color CCD camera over the production line helps to obtain the picture of layer of the stack in real time. The obtained image will then experience HSV picture segmentation to extract the standard titanium dioxide labels out for detection. Multiple morphological processes are applied to transform the labels into detectable rectangles. After that, the angle of inclination of each package would be detected and calculated through Hough transform. A tolerance range is set to distinguish qualified stacks and unqualified stacks by the difference between the detected angle and standard angle. This detection system could significantly improve the efficiency of production.