A tri-layer window in time domain is proposed, and a multiscale temporal relative local contrast measure (TRLCM) is developed to detect infrared (IR) small moving targets. The proposed algorithm divided some consecutive frames into five parts: the peak, the two guards, and the two baselines. By adjusting the width of the five parts of the tri-layer temporal window, the TRLCM can achieve the multiscale detection in time domain, so it can successfully detect the targets of different sizes and speeds. Besides, both ratio and difference operations are utilized to calculate the temporal local contrast information, so that the true target can be effectively enhanced and complex background can be suppressed simultaneously. Experiments for seven true IR sequences show that the method can effectively extract the small moving targets under complex background, and, comparing to some existing algorithms, the method can achieve a better performance in detecting rate and false alarm rate.
Small Moving Target Detection in Infrared Sequences by Using the Multiscale Temporal Relative Local Contrast
Lect. Notes Electrical Eng.
2021-10-30
13 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Temporal Profile Based Small Moving Target Detection Algorithm in Infrared Image Sequences
British Library Online Contents | 2007
|