A Study On The Algorithm Of Ship Target Detection In SAR Imagery And Its Applications

In recent years, ship detection and surveillance in Synthetic Aperture Radar (SAR) images has received considerable attentions in the area of marine remote sensing. This issue has been one of the most important marine applications of SAR data. The study on automatic detection algorithm is urgently needed when great amount of SAR images are used in ship target detection. This study can extract information of ship targets quickly and effectively.This paper, which background is engineering application, is based on the theory of the understanding of SAR image. It studies on the aspects of ship target detection, ship wake detection, ship feature extraction and establishment of the Sea Area Target Surveillance System.The state-of-the-art of the study of ship and ship wake detection is summarized. The target types of the ship and its wake in different resolution SAR images are analyzed and a conclusion is draw. The issue of ship target detection to the low or middle resolution SAR image is point target detection while the issue to the high-resolution SAR image is hard target detection. The issue of wake detection is linear target detection.The existing algorithms of ship target detection are summarized. An algorithm of local window based K-distribution CFAR is proposed. Using experimental results, the traits and the fitness scopes are analyzed for the two-parameter CFAR algorithm, the K-distribution CFAR algorithm and the local window based K-distribution CFAR algorithm. On the other hand, the characteristics of ship target in high-resolution SAR image are analyzed. The KSW double thresholding method is proposed to ship detection in the high-resolution SAR image.In the area of wake detection algorithm, the deficiency of the detection algorithm based on RADON transform is pointed out. A wake detection algorithm based on normalized grey level HOUGH transform is proposed. It is improved to normalized grey level based on the standard HOUGH transform. The calculation expression of the wake end points is derived.According to the characteristics of the ship targets in SAR image, it is proposed that there are three types of ship feature. They are geometry feature, geographic feature and kineticfeature. The parameter calculation methods pre also given. A ship target detection system, which based on the study of target detection, wake detection and feature extraction, has been developed.Finally, the functional design and key technologies of the Sea Area Target Surveillance System for Remote Sensing Images are discussed. The differences of the ship targets between the optics image and SAR image are analyzed primarily. The research of this paper has accomplished the main study work of the system kernel. It guarantees the realization of the key technology.