An evolutionary multi-agent system for object recognition in satellite images

Abstract

This paper proposes a new approach to combine the knowledge-based model and the cooperation technique of evolutionary agents to identify the location of the desired object in a satellite image. The agents interact with the local information of the image pixels to search for the target objects through an evolutionary process. A new set of fitness function and evolutionary operators are defined for the process. The decentralized, bottom-up and evolutionary natures of the agents can be used to construct a robust system for object recognition in satellite images. The experimental results are satisfactory and have demonstrated the flexibility and power of the approach. © 2003 IEEE.