Developing stereo vision system for object recognition and positioning of AMAD-R mobile robot

Abstract

Machine vision can be described as a useful robotic sensor since it mimics the human sense of vision and allows for non-contact measurement of the environment. The infinite number of possible poses relative to the viewer will cause the3-D object gives rise to an infinite variety of 2D images or views. Two things must be known—“what” the object is and “where” it is located i.e. the object has to be recognized and its coordinates must be known, In order for a vision system to be effective in assisting a robot to approach an object autonomously. This research presents a type of stable target acquisition system for mobile robot. A combination of vision sensors and mobile robot will be use in order to measure accurately the location of a target in real world coordinates. Then, a robust target extraction algorithm that able to pick out a target in an image will develop. After this is done we will develop a stereo vision algorithm to find the distance from the robot to the target. Finally we will integrate all of these algorithms together. This is important in order to develop a target extraction technique that will be able to find the distance accurately to the target. With the distance we can then locate the real world position of the target.