Dense Stereo Correspondence Algorithm for Robotic Applications
Deepambika V.A1, Arun lal S.L.2
1Deepambika V. A, Department of Electronics and Communication, LBS Institute of Science and Technology for Women, Trivandrum, (Kerala), India.
2Arun Lal S.L, Scientist, ISRO, Thiruvananthapuram, (Kerala), India.
Manuscript received on September 30, 2013. | Revised Manuscript received on October 10, 2013. | Manuscript published on October 30, 2013. | PP: 439-442 | Volume-3, Issue-1, October 2013. | Retrieval Number: A4470103113/2013©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Stereo vision, the passive sensing technique for inferring the three dimensional position of objects of a scene under study is having great applications in the field of machine vision, robotics, image analysis and image reconstruction. Robotics require computationally fast and easy to implement stereo vision algorithms that will provide reliable and accurate results under real time constraints. By using some similarity measure, the stereo correspondence, tries to find out the matching pixels or objects between left and right views of the scene. Since the focus is on real time application the local winner-take-all optimization in the disparity computation process is done in this study. The correspondence is done by using fast block matching Sum of Absolute Differences (SAD) algorithm .With the help of camera parameters and the disparity map obtained from this algorithm, the depth map of the scene under study is extracted by using the principle of triangulation. To simplify the correspondence search, rectified stereo image pairs are used as inputs.
Keywords: Stereo Correspondence, Sum of Absolute Differences (SAD), Disparity, Depth.