Optical Flow Vectors Thresholding in Assisting Heading Direction Estimation

Abstract

This paper presents a technique on how to estimate heading direction of a moving vision system such as mobile robot. The heading direction is represented by estimation angle which is generated by optical flow vectors threshold technique (OFVTT). The utilizations of optical flow field generated based on Horn-Schunck and Lucas-Kanade methods are essential in order to compute the threshold value. The performance of the proposed technique was determined through percentage of root mean square error (RMSE). Based on our experimental results, it can be ascertained that combination of Horn-Schunck method together with OFVTT shows the best performance in assisting the heading direction estimation of a moving vision system.