Abstract
In the industrial image processing, a template matching method search a target image area in an objective image area. However, failure often happens when the part of the object is occluded. This paper proposes a method to search a target image area even if a wide part of the objective image is occluded. In the proposed method, template image is separated into multiple block areas and each block area is matched with an objective image. The result is brought by the unification of the result of each block. The weight of block is based on the matching rate of the second solution for a reference image. The weight of each block area is revised at the collation coordinate of other block areas. Influence of occlusion is reduced by this processing. Therefore, this method is robust to the occlusion. The experimental results show that the proposed method is more robust than conventional method to occlusion.
