Small Infrared Object Detection Method Based on Cycle Spinning Contourlet Transform
基于循环平移Contourlet变换的红外小目标检测方法

The methods for detecting small infrared targets in the case of noise and background interference are analyzed. A small infrared target detection algorithm which combines the cycle spinning contourlet transform with the adaptive threshold segmentation is proposed. Firstly, the method uses the cycle spinning to de-noise the original images. Secondly, it obtains the residual images by subtracting the de-noised image from the original image. Thirdly, it uses the adaptive threshold segmentation to segment the residual images so as to separate a few of candidate target points. Finally, it uses the continuity and consistency of target motion to detect the target. The simulation result shows that the proposed detection method can detect the small infrared targets in the sequential images precisely and it is more effective than the Contourlet transform and wavelet transform methods in detection.