An Object Detection Algorithm with Disparity Values

It is a technically difficult issue that the object segmentation under real world applications, because of the overlapping problem and the scattering disparities of the same object. This paper proposes an object detection algorithm with disparity values to solve the problem. The proposed algorithm uses the connected component algorithm and disparity values to connect pixels with similar disparities, and the recursive object merge uses to merge the similar neighboring contours for locating and bounding the contours of an object. Experimental results show that the proposed algorithm can resolve the problem of object detection in overlapping objects, and the detection results can be used for subsequent object recognition applications.