Vision Recognition System by Using Chaotic Search

Abstract

This research is concerned with image recognition for a robot vision detecting target objects by using Chaotic Search in model-based matching. As a nonlinear dynamical system to generate a chaos, BVP model is treated, which shows the behavior of neurons in biological system. This model has the "edge of chaos", which exists on the boundary between a periodic solution and a chaos solution. This edge of chaos is an important area to maintain an organization to be flexible. In this research, the Chaotic Search is applied to image recognition utilizing the edge of chaos. First, the occurrence of chaos in BVP model is examined using the Lyapunov exponent. Second, in order to perform image recognition, we propose a method of Chaotic Search in which it can distinguish the target object from surroundings effectively, using a method of pattern matching. Finally, through two illustrative examples, the effectiveness of Chaotic Search is verified for both static and dynamic targets.