Infrared Small And Weak Target Detection Algorithm Based On Local Contrast Measurement

Infrared imaging is widely used in target detection fields due to its advantages of long detection distance,strong concealment,and all-weather work and so on.Weak target detection is the key technology for infrared guidance and the prerequisite for target tracking which deserves important research significance.Some typical scenes usually contain strong background noise and complex cloudy interference,and the target itself occupies fewer pixels,which makes the traditional detection algorithm less satisfactory in accuracy and robustness.This thesis focuses on practical problems such as dim targets which with low brightness,or complex highlighten background and pixel-level hlighten noise interference,then makes research on local contrast measurement and image entropy to make up the shortcomings of existing algorithms for processing images with low signal-to-clutter ratio.The main work is summarized as follows:1.Aiming at the shortcomings of low detection rate,poor robustness,insufficient applicability and high algorithm complexity of the existing infrared small and weak target detection algorithms when processing low signal-to-clutter ratio images,this thesis proposes a multi-scale local contrast mehod for infrared small and weak target detection based on human visual mechanism.This algorithm fully considers the gray difference between the target and the background area,and uses the visual contrast mechanism to improve the visual saliency of the small infrared target in the low signal-to-clutter ratio image.First,defined a newly local contrast operator to enhance the contrast of the original image;then,the multi-scale method is used to optimize the salient regions of the image,so as to increase the applicability of the algorithm for different sized-targets;finally,the adaptive threshold segmentation method is applied to obtain the final target which to be detected.The proposed algorithm uses the strategy of enhancing the visual saliency of the target area to realize target detection,it is also robust to background clutter and pixel-level noise,furthermore,it does not require image preprocessing operation.All those show the advantages of the algorithm as simple,strong real-time performance and easy engineering applization.Experimental results reveal that the proposed algorithm can effectively detect different sized weaktargets.Compared with existing algorithms,the performance of the target detection has been significantly improved.2.Aiming at the problem that the local contrast algorithm has a target block effect at a large scale when enhance the target region,which causes a deviation in the detection result,then an infrared weak and small target detection algorithm based on entropy-weighted local region contrast is proposed.By designing the traversal template and using the image entropy characteristics,the algorithm effectively improves the visual saliency of the targets and suppresses the generation of target blockiness.First,the newly defined local region contrast is used to enhance the target and suppress the background to improve the visual saliency of the target;then,the local entropy operator is used to suppress the interference of complex edges in the detection;finally,the adaptive threshold segmentation method is applied to obtain the real target to be detected.The proposed algorithm can effectively enhance weak and small targets without image preprocessing.It has the advantages of easy implementation.Compared with existing algorithms,the detection rate and robustness of the algorithm are intuitively improved.