Transformer-based segmentation network for false target suppression in radar detection

During the process of radar detection, various factors such as noise and clutter interfere with the received echo signals, leading to the detection of a large number of false targets. This poses significant challenges to the detection, tracking, and identification of real targets. This paper introduces a Transformer-based segmentation network for false target suppression. It innovatively transforms the problem of false target suppression in sequential signals into an image segmentation task. The dice loss is effectively employed to address the issue of class imbalance between targets and backgrounds. Furthermore, the introduction of the transformer module further enhances the segmentation performance of the model. The proposed method is validated on real radar measurement data, demonstrating both the effectiveness of the designed model and the improvements brought by each module.