An Improved Multiscale Patch-Based Contrast Measure for Small Infrared Target Detection

Abstract

Aiming at the problems of insufficient background suppression and weak multi-target detection abilities of existing infrared small target detection methods, which lead to high false alarm rate and high omission factor of infrared search and track system, an infrared small target detection method fusing modified anisotropic diffusion coefficients with multiscale patch-based contrast measure (ADMPCM) was proposed. The local contrast values of four different directions in the local area are applied into the modified anisotropic diffusion coefficient equation, and the final filtering result is the minimum function value of the four equations. Extraordinary experimental results revealed that, in average, background suppression factor increased 2.95 times, signal-to-clutter ratio gain increased 6.17 times on single-target detection task and 10.49 times on multi-target detection task, respectively, compared with similar detection methods.