A two-camera-based vision system for image feature identification, feature tracking and distance measurement by a mobile robot

Abstract

This paper presents a two-camera-based vision system for image feature selection, tracking of the selected features and the calculation of 3D distance of the selected features. The feature tracking approach is based on minimisation of the sum of squared intensity differences between the past and the current window, which determines whether a current window is a warped version of the past window. The 3D positions of these features can be calculated on the basis of the known image coordinates of the same point/window in the left and right camera images. The distance calculation is carried out by employing ‘midpoint of closest approach’. The vision system with the controlling architecture is implemented with the KOALA mobile robot. The system has been tested for real life environment in our laboratory and the experiments showed that the system can reliably detect features and track in subsequent frames and the 3D distances calculated for tracked features showed satisfactory accuracy.