Superpixel-Based LCM Detector for Faint Ships Hidden in Strong Noise Background SAR Imagery

Abstract

Faint or small target detection is always a challenging task for image-based remote sensing application. Due to low signal-to-clutter ratio, many conventional methods fail to yield satisfactory results. To solve the problem, a superpixel-based local contrast measure (SLCM) detector is proposed against the faint ship targets hidden in strong noise background synthetic aperture radar (SAR) images. The detection stage consists of superpixel segmentation and SLCM detection. In the first stage, a modified simple linear iterative clustering algorithm is utilized to segment SAR image into superpixels. In the second stage, a novel SLCM is performed on the generated superpixels to enhance and find the hidden ships. Experiments on real SAR images indicate that the proposed method can effectively detect ship targets from the strong noise background SAR images.