Study On The Methods Of Ship Target Detection In SAR Imagey

Recently, ship detection on Synthetic Aperture Radar(SAR) imagery is an important way of ocean surveillance and fishing monitoring. China has large ocean which is full of resources. It is of grate significance in developing ship detection researchment on SAR imagery. This paper focuses on the sea clutter modeling,sea-land segmentation in SAR image and the algorithms of Constant False Alarm Rate(CFAR) in SAR imagery ship detection. The main contributions of this paper can be summarized as follows:1. Because accurate modeling of sea clutter affects the performance of CFAR based on statistical models. This paper studies the statistical characteristics of SAR sea clutter,presents the setup of statistical modeling and the criterions of model fit performance,and assesses several models’ modeling performances to sea clutter through experiments.2.Sea-land segmentation is a basic of the SAR vessel detection, analyzed and discussed the method of Sea-land segmentation.On the basis of gray level-average gray level-median gray level three-dimensional(3-D) histogram, put forward gray level-average gradient-median gray level 3-D histogram, Based on gray level-average gradient-median gray level 3-D histogram, the improved regional division and fast iterative formulas of 3-D Otsu’s thresholding method are derived. The Sea-land segmentation experiment show that the proposed algorithm has accurate for segmentation and reduce the time.3. For the ship detection in medium and high resolution SAR imagery, an improved CFAR detection algorithm based on G0 distribution is proposed in this paper. Firstly, according to the contrast parameter of the sliding window to initially determine whether the ship pixels within the sliding window, continued follow-up of CFAR detection if the sliding window has ship pixel.Otherwise sliding window moving in steps of half a window instead of a pixel, it can reduce the number of time-consuming parameter estimation. The experimental results show that the method has a good performance and reduce the running time.4. A software system for SAR image ship target detection is designed. Introduces the functional modules of the system include SAR image pre-processing module, sea clutter distribution fitting test module and ship detection module.