Feature parameters selection based on rough set theory for target recognition

Abstract

Rough set theory was applied to select optimal feature parameters of target recognition systems in this paper. Target recognition systems were treated as a kind of information systems with decision. Firstly, the original crude information caught by sensors was transformed to information table. Then, discernable matrix method was adopted to simplify the table and delete redundant attributes and recognition rules. Finally, the optimal feature parameters group was extracted, and the simple and accurate rules for target recognition were obtained. The validity and advantage of this approach were demonstrated by an example.