In this paper, we examine the use of image segmentation approaches for target detection in TWRI. The between-class variance thresholding, entropy-based segmentation, and K-means clustering are applied to segment target and clutter regions. Real 2D polarimetric images are used to demonstrate that simple histogram-based segmentation methods produce either comparable or improved performance over the Likelihood Ratio Tests (LRT) detector. Specifically, the results show that, for the cases considered, the entropy-based segmentation outperforms the other image segmentation methods and the LRT detector.
展开▼
