Infrared Dim and Small Targets Detection Based on Multi-scale Local Contrast Measure Utilizing Efficient Spatial Filters

Given the growing importance of infrared search and track (IRST) systems, in this paper, inspired by the mechanism of the human visual system (HVS), we present a robust algorithm for the detection of infrared (IR) dim and small targets based on Multi-scale Local Contrast Measure utilizing Efficient Spatial Filters, which is named as ESFMLCM. In the proposed algorithm by considering a nested structure, three efficient spatial filters, namely Gaussian filter, mean filter, and standard deviation filter, are applied to each pixel of the raw IR image, simultaneously. According to the idea of the matched filter, the Gaussian filter has been used to improve the actual targets and increase the detection rate. Also, two other filters have been used to suppress all the types of interferences (noise and background clutters) and reduce the false alarm rate. Finally, a simple adaptive threshold operation is applied to extract actual targets or remove fake targets. The numerical and visual results show that the proposed algorithm can detect IR dim and small targets under complex backgrounds with high accuracy in comparison with other algorithms.