Infrared Dim And Small Target Detection In Complex Background

By the comprehensive and in-depth studies of the low Signal-to-Noise(SNR) characteristics of the infrared image on the complex background and based on small target detection algorithms both home and abroad, the paper presents several effective infrared dim and small targets detection algorithms.The first method, for the study of complex background and the real-time requirements, comprehensive background suppression is finished by fusing Laplace high-pass filtering and median filtering algorithm. At first, the majority of background filtering uses the high-pass filter, then uses the median filter to smooth image and filter noise interference, finally real targets is detected by using the adaptive threshold segmentation method.The second method uses improved mathematical morphology filter for background suppression. The method mainly adopts double-suppression for background clutter. First, the optimal structural elements are selected for making preliminary background suppression, and then difference results is computed; further, the second background suppression is adopted and last the corresponding threshold segmentation is adopted to detect the targets signal.For infrared sequential images, the third method, according to targets characteristics of inter-frame correlation, is to detect targets coordinates location by using the 8-neighborhood detection algorithm.A lot of experimental results demonstrate that several methods presented infrared dim and small targets detection in the complex background are very effective, exact.