Research On Infrared Small Target Detection Method In Sea-Sky Background

The infrared detection of small targets in the sea-sky background has been a hot and difficult point in the field of image processing.Due to the small target has no obvious characteristics,small size,small number of pixels and so on,it is more difficult to detect the small target from a long distance,and it is difficult to obtain the related information of the small target.At the same time,due to interference factors such as sea clutter,sea antenna and sky clouds,the target is easily submerged in background noise and clutter.Aiming at the detection method of infrared slow small target in complex sea-sky background,the research contents and innovations of this paper are as follows:Aiming at the difficulty of accurate target detection in the background of high sea,the preprocessing algorithm of regional target detection is studied.By analyzing the statistical distribution model of the infrared image area,and according to the difference of the regional distribution model,the target area that meets the requirements is selected.In order to verify the feasibility of the algorithm,infrared small target detection is carried out in the sea-sky background.The results show that the preprocessing algorithm of regional target detection can achieve the initial positioning of the target,which is of great significance to improve the detection rate of small target,reduce the false alarm rate and the missed alarm rate.Aiming at the elimination of clutter and noise points in complex sea-sky background and the adaptive detection of targets in different sea-sky scenes,an adaptive target segmentation algorithm based on regional target detection is studied.A mathematical model is established to get adaptive thresholds of different backgrounds,and then the binary image of the target region is obtained according to the threshold segmentation,and the region image is reconstructed to obtain the detection results.The simulation experiments of single target,multi-target and single frame infrared images in different sea-sky backgrounds are carried out respectively,and the results show that the algorithm has better detection effect for single target,multi-target and different sea-sky backgrounds.In order to determine the real target,an improved pipeline filtering algorithm is introduced to synchronize the detection and counting of the target in the sequence image.A combined algorithm based on ATS-RTD and pipeline filtering is proposed to verify the two groups of sequence images.The results show that the combined algorithm can accurately detect the real small target and has strong anti-interference ability to the background.Aiming at the real-time performance of infrared small target detection in sea-sky background,adaptive and stable detection of targets in different sea-sky scenes,the block mean algorithm based on area target detection is studied.The block matching template is established for the preprocessed infrared image,and the target in the region is obtained by matching with each pixel in the target region.The final detection result is obtained by region image reconstruction.The simulation results of three different infrared images show that the algorithm has.good detection performance for single target,multi-target and different sea-sky background.Compared with Top-hat,Otsu and other four algorithms,the results show that the algorithm has the advantages of high real-time,adaptive and high detection rate.In order to eliminate noise points and other interferences,a small target detection algorithm based on BM-RTD and pipeline filtering is proposed,which is verified by experiments on two sets of sequence images.The results show that the combined algorithm has good real-time performance and can be applied to the detection of small targets in different sea-sky backgrounds,with high detection rate and anti-interference performance.