Infrared Small Target Detection Algorithm Based On Local Contrast Enhancement

As an important part of the military target detection,early warning and tracking system,infrared small target detection technology has been widely concerned by scholars.Due to the long imaging distance,the target is often presented in spots in the infrared image,lacking clear structure and texture features.At the same time,due to the noises from transmission and the background interference,the target is often submerged in the background and noises.Therefore,it has a great theoretical significance and application value to study infrared small target detection under complex scenes.According to different scenarios of the infrared small target detection problem,this paper has designed two kinds of infrared small target detection algorithm based on local contrast enhancement,which are infrared small target detection algorithm based on LoG fusion and local contrast enhancement and infrared small target detection algorithm based on multiscale local homogeneity measure.The main research content is as follows:(1)An infrared small target detection algorithm based on LoG fusion and local contrast enhancement is designed.The algorithm uses the central excitation and lateral suppression characteristics of the LoG filter and the negative LoG filter to obtain the circular spot structure information in the infrared image respectively,and realizes the simultaneous response to the bright target and the dark target in the same picture.On this basis,combining with local contrast methods,the target is further enhanced while background and noise are suppressed.Experiments show that the algorithm has good detection performance and achieves stable and rapid detection in different scenarios.(2)An infrared small target detection algorithm based on multiscale local homogeneity measure is designed.The complex background edges in the infrared image also tend to produce local contrast,which brings interference to the real infrared small target.In order to distinguish the small target area from the normal background edge,the algorithm uses local contrast information with directionality as the basis for the detection of infrared small targets.At the same time,combining the characteristics of high homogeneity of gray values in small target areas to achieve the target enhancement as well as to background and noise suppression.Experiments on three different scenes show that the algorithm has good ability on target enhancement and background suppression,and has strong anti-noise performance.