A New Straight Line Matching Technique by Integration of Vision-based Image Processing

Abstract

In this paper, a new algorithm for a new straight line matching by integration of vision based image processing is being proposed. The corner detector and Hough Line Transform were employed respectively the feature extraction of corners and edges from the left and right images. After process of corner detection has been already extracted the corner points in the left and right images, each corner use the computation of a cross-correlation which is based on matching score for putative corner correspondences. This algorithm uses the corner pairs from reference points of started-end corner. The matching comparison between corner pairs and Hough Line are able to generate a new straight on the edge of both left and right images. From corner correspondences in the image pairs are going to compare with a new straight line to get the result of new straight line correspondences. This method is simple and easy to apply to the 3D recognition of geometric shapes and is able to reduce the number of false matches in a probabilistic manner which the result arise from incomplete Hough Lines. The results have shown the new straight line correspondences that are effective, complete and could be used in a 3D recognition process.