An infrared-small-target detection method in compressed sensing domain based on local segment contrast measure

Real-time performance is one of the key properties in infrared targets detection system which limits the applications of many algorithms. In this paper, a novel infrared-small-target detection method using compressed sensing technology is proposed to improve real-time performance by combining the images compressed and targets detection. Furthermore, dealing with images with both bright targets and dark targets, filter images for two kinds of targets separately is commonly for the image preprocessing. In this paper, a local segment contrast measure method is proposed to preprocess images uniform. Finally, the influence of certain vital parameters (e.g., the block size and the filter window size) on the detection and compression performance is discussed at length. Several guiding principles for the selection of those vital parameters are developed. The experimental results demonstrate that the proposed framework with an appropriate block size and filter window size provides a great balance between real-time performance and accuracy. The proposed local segment contrast measure method is efficient when applied to both bright and dark small targets.