RADAR TARGET HRRP SEGMENTATION USING DIFFUSION MODEL

High-resolution range profile (HRRP) formed by the extended target under high-resolution radar can reflect the detailed information of the target, which is one of the important means of target detection for air-to-ground radar. Accurately segmenting the target HRRP from the entire echo is of great significance for studying the target characteristics, target feature extraction and recognition. In this paper, a radar target HRRP segmentation algorithm using diffusion model is proposed. The disturbed position parameters of the target HRRP are generated according to the idea of adding noise by the diffusion model, and then the predicted segmentation results are obtained through the trained model to modify the disturbed parameters. The performance of the algorithm under different signal-to-noise ratios and polarization modes is evaluated in a measured dataset. The experimental results show that this method can effectively obtain target HRRPs and can be applied to the extended target segmentation of high-resolution radar.