Research On Detection Algorithm Of Small Target In Radar Image

With the development of economy and society,the implementation of the strategy of a country with strong transportation network,the demand of intelligent shipping,intelligent port,intelligent supervision in the shipping industry are rising.Traditional radar,CCTV has been unable to meet the needs of monitoring and supervision.Synthetic aperture radar not only has the advantages of all-weather,all-day,but also has the feature of high resolution,active.Detection of small targets at sea has always been a difficult problem in the field of detection and tracking of targets.At present,SAR images are used as an important tool for detection of small targets at sea.This paper mainly focuses on the research of small target detection in SAR images.Constant false alarm rate algorithm is the most classical method in the field of small target detection.This method relies on the contrast between background and target and the model of background clutter.With the improvement of image resolution,the failure of distribution model leads to the decrease of detection performance.In addition,because of the small and weak target has weak radar echo which caused the target can not be detected.In view of the above problems,this paper studies two aspects:image denoising algorithm enhances the contrast between the background and the target,and then separates the background to the maximum extent for accurate modelingFirstly,this paper describes the theoretical basis of image small object detection in SAR images.The factors affecting the visibility of small targets are described,and the basic ideas of this paper are put forward according to the current situation of detection of small targets in images.Then,in view of the fact that the traditional denoising algorithm can not solve the problem of "point singularity" and "line singularity" at the same time,which will cause excessive denoising.A novel fusion denoising algorithm based on adaptive threshold and curved wave transform is proposed based on wavelet coefficients in different subbands.The adaptive threshold and the enhanced curved wave algorithm achieve approximation of "point" and "line" perfectely,and the effect is better.Next,according to the characteristics of high resolution of SAR images,the Gaussian and non-Gaussian models for background clutter are introduced.In order to find the most suitable model,parameter estimation is determined by using the maximum entropy principle to segment the background.The χ2 and KS distance are introduced to test the goodness of fit.Finally,based on the problem of low detection performance of traditional CFAR detectors,a CFAR detection algorithm based on the principle of maximum entropy is proposed,which solve the problem of high false alarm rate and low detection performance.The experimental results show that the proposed fusion denoising algorithm can enhance the contrast between the background and the target on the basis of smoothing the background clutter,and the performance of the proposed detection algorithm is improved.