Stereo Vision for Visual Object Tracking and Distance Measurement Assessment

Abstract

The computer vision for binocular eyes system has many applications in robot applications and safety purposes. Based on the previous research, the combination of the area of sight of stereo vision will trigger the trigonometry intersection point for determine the distance of the objects from it baseline. The system programme codes is one of the issue need to confront consequent since there are various sorts of calculation that are in the same field, however has unmistakable of use. This project focuses on how to measure distances using binocular vision. The main objective is to evaluate the binocular vision system by calculating the distance of objects in real environment. Furthermore, the project proposes a new program algorithm for binocular vision system to work, in order to identify distance of an object with a basic equation has been derived and set in the designed algorithm. The setting environment are set to single and multi-objects measured, object in environment and changes of degree of bright light. Evaluation of the system shows the detected distances are consistence and the data were recorded. The value of the distances detected are then compared with the real environment distances. The result show distances measured moderate enough for proposed system to function and may facilitate improvements in computer vision system for industry.