Infrared Small Target Detection via Center-Surround Gray Difference Measure With Local Image Block Analysis

Exristing algorithms may suffer from high false alarm rate and low detection probability when detecting dim small target under intricate background clutters and heavy noise. To address this problem, a target detection method based on local image block analysis and center-surround gray difference measure (CGDM) is proposed in this article. First, the infrared image is decomposed into a group of local blocks. The gray information of each block are sorted in ascending order and the last element is deleted, then the maximum gray jump point (MGJP) is extracted by using differential operation. Second, a protection zone is formed with MGJP as the center, and the maximal intensity of protection zone (MIPZ) is extracted. Next, by comparing the gray intensity of MGJP and MIPZ values, the potential target blocks can be reliably selected, while most background blocks and noise blocks will be discarded. After that, the CGDM is presented to enhance the target and suppress background clutters, and then, the center-surround local contrast measure (CLCM) is designed to further suppress the high-intensity clutter residues by searching the target center and revising the protection zone. Finally, the weighted center-surround gray difference measure (WCGDM) is defined by CLCM WCGDM map to recognize real targets. Extensive experiments show that the proposed method outperforms several existing algorithms in small target detection under complex background clutters, and it is robust to various target shapes, target sizes, and noise types.