Ship Detection Algorithm Research Of High Resolution Optical Remote Sensing Image

In this paper, engineering background, combined with imaging characteristics of the optical images, using image processing and pattern recognition technology to detect the core target ship launched a series of research work. Discussion on the ship target detection several key issues.Currently optical remote sensing image ship target detection algorithm has many shortcomings, they are mostly the detection rate, false alarm rate, robustness, high time complexity is not conducive to engineering and other aspects, it has seriously hampered the General Ship development of the detection system. Based on this phenomenon we propose a practical ship target detection algorithm. Algorithms are separated by the sea, target segmentation, target detection of three modules.The current sea separation algorithm mainly threshold method based on reflectivity,when an area of land and marine areas close reflectance or vice versa, is likely to cause an error segmentation. In response to this shortage, land and sea image separation module introduced entropy feature, search the image entropy seeds, were homogeneous region growing, greatly reduces the probability of erroneous segmentation, segmentation robustness improved significantly. Most of the traditional ship target image segmentation algorithm to analyze the spatial domain, common spatial domain processing methods can not achieve the desired segmentation. This paper presents transform analysis method will be introduced based on the Fourier spectrum residual significance ship target segmentation calculation method to solve the problem of poor in the spatial domain segmentation results,in particular to reduce the distribution of broken cloud brought influences.Finally, a plurality of feature amounts, taking into account the whole algorithm can be realized, ship target detection module uses voting method to extract segmentation results,one-out target false alarm results show that the optical remote sensing image ship detection algorithm proposed in By holding 12.71 percent of the false alarm rate for achieving adetection rate of 90.32%. Single 8192 x 8192 pixel image test with a time of 1.25 seconds,better than the current ship detection algorithm. Suitable for use in ship detection system.