A real-time middle wavelength infrared multi-target detection and tracking algorithm based on LoG

This paper presents a real-time middle wavelength infrared multi-target detection and tracking algorithm based on LoG (Laplace of Gaussian) to resolve the poor stability and real-time of some infrared multi-target detection and tracking algorithms. According to the features of middle wavelength infrared images, this paper uses the scale invariance of LoG to detect multi-target and it can track targets in real-time. Firstly, this paper analysis the features of middle wavelength images, then uses the features of targets and backgrounds to segment the targets from backgrounds. Secondly, this paper uses LoG operator to detect multi-target. Lastly, this paper computes horizontal and vertical projection histograms of targets in the tracking windows and computes the regional connectivity of histograms to track targets. This paper uses infrared target detection and tracking system to test this algorithm, the results show that this algorithm can detect and track multi-target of different scales stably in real-time.