VISION SYSTEM FOR DEPALLETIZING ROBOT USING GENETIC LABELING

In this paper, we present a vision system for a depalletizing robot wh ich recognizes carton objects. The algorithm consists of the extractio n of object candidates and a labeling process to determine whether or not they actually exist. We consider this labeling a combinatorial opt imization of labels, we propose a new labeling method applying Genetic Algorithm (GA). GA is an effective optimization method, but it has be en inapplicable to real industrial systems because of its processing t ime and difficulty of finding the global optimum solution. We have sol ved these problems by using the following guidelines for designing GA: (1) encoding high-level information to chromosomes, such as the exist ence of object candidates; (2) proposing effective coding method and g enetic operations based on the building block hypothesis; and (3) prep aring a support procedure in the vision system for compensating for th e mis-recognition caused by the pseudo optimum solution in labeling. H ere, the hypothesis says that a better solution can be generated by co mbining parts of good solutions. In our problem, it is expected that a global desirable image interpretation can be obtained by combining su b-images interpreted consistently. Through real image experiments, we have proven that the reliability of the vision system we have proposed is more than 98% and the recognition speed is 5 seconds/image, which is practical enough for the real-time robot task.