Detection and tracking of small infrared motion targets using directional Top-hat algorithm

Abstract

Small targets are shown on the infrared image as small bright spots, do not carry any characteristics of the target, are difficult to detect, but are the object of reconnaissance and detection equipment, according to long-range tracking. On the basis of the characteristics of the small infrared target and the Top-hat algorithm, a method of using the directional Top-hat filter is proposed to remove the background and enhance the target. Taking advantage of the rich information in consecutive shots of a video and the slow change in position between frames, using the search method from the vicinity of the target position in the previous frame to determine the target, thereby limiting the missing object being tracked. The test results show that the directional Top-hat algorithm is effective in detecting small targets in complex backgrounds, combined with the target recognition algorithm of consecutive frames, allowing precise tracking of the target of interest.