Infrared Small Target Detection Based on Sub-maximum Filtering and Local Intensity Weighted Gradient Measure

The development of infrared search and tracking (IRST) systems owes a great deal to the crucial role played by infrared small target detection technology. However, the interferences caused by pixel-sized noises with high brightness (PNHB) and high-intensity structures in background regions make it a tough task to realize robust detection of low-contrast targets. To this end, we propose an infrared small target detection method named sub-maximum filtering and local intensity weighted gradient measure (SF-LIWGM) in this paper. First, a maximum eight-neighborhood sub-maximum (MENS) filter is designed to eliminate the PNHB without damaging the structure of the small target. Besides, we construct a novel local measure framework by aggregating the intensity and gradient information, in which the detection process is split into two branches: target saliency enhancement and high-intensity structural clutters (HISC) suppression. Finally, we adopt an adaptive segmentation operation to implement the target detection. The experimental results on real infrared data demonstrate that our method has satisfactory performance against other state-of-the-art algorithms in terms of signal-to-clutter ratio gain (SCRG) and background suppression factor (BSF), with an average improvement of 5.13 and 4.16, respectively.