Abstract: The time-frequency joint distribution of a target reflects well its physical structural property so it can be used as an effective means of radar target recognition. Aiming at the problems of low identification rate and poor antinoise performance in existing time-frequency analysis methods, a method of radar target electromagnetic features extraction and recognition based on S-transform is proposed. The time-frequency distribution of a target is first analyzed by the S-transform and the stability of time-frequency distribution within a large angle range and signal-noise ratio is derived. Based on these stable signatures, a recognition algorithm is proposed. Finally, the identification of different target models are performed and the results demonstrate that the method provided in this article has higher recognition performance and better robustness than other time-frequency-analysis methods.

Key words: information processing, feature extraction, S-transform, target recognition, time-frequency analysis