WCSE

Narrowband RCS Feature Selection and Evaluation in Applications

Abstract— Narrowband RCS is abundant in target characteristic information, and selection and evaluation of characteristic directly affect target identification. Sliding window is used for feature extraction in engineering applications. Window size and sliding step are the two basic parameters of sliding window method. Fisher criterion is adopted to estimate the effect of the parameters in narrowband RCS feature separability. Experimental results show that both the number of separable features and the feature separability increase,as the window size and sliding step increase. With different sliding window parameters, RCS spectrum mean which reflects the average distribution of spectrum exhibits good separability for the classification of aircraft and first stage.

Index Terms— feature selection, sliding window parameters, Fisher criterion, feature evaluation.

Yu Cheng
Jiuquan Satellite Launch Center
Xi Li
Jiuquan Satellite Launch Center
Lijun Liu
Jiuquan Satellite Launch Center

[Download]

Cite: Yu Cheng, Xi Li, Lijun Liu, " Narrowband RCS Feature Selection and Evaluation in Applications, " WCSE 2022 Spring Event: 2022 9th International Conference on Industrial Engineering and Applications, pp. 1515-1520, Sanya, China, April 15-18, 2022.