HOUGH变换在海浪遥感图像处理中的应用

Application of HOUGH Transform in Analyzing Ocean Wave From Seasat SAR Image

本文提出了一种以FIOUGH变换为基础的海浪波长波向分析法,并对SEASAT合成孔径雷达获取的海洋遥感图像进行了方向和波长分析。该方法模拟人类纹理视觉机理,首先应用CFCM动态聚类方法对波峰、波谷和过渡带进行分类,然后采用随机行走模型对分类图像进行滤波,以获得整图的全局走向信息,再以此先验信息为先导对分类后的纹理图像进行定向跟踪拟合,得到代表海浪波峰的直线族图,最后在再对此图进行HOUGH变换和分析的基础上得到精确的波长与波向。

A new method based on HOUGH transform for analyzing ocean wave from SEAS AT SAR image is presented here. This method takes the way which human vision takes to analysis nature texture image, after extracting the peak area from the SEASAT SAR image with CFCM classification method, we use random-go model to filter the classified image in order to obtain the global direction information of sea wave in the whole image. Then, according to this information, a directional tracking algorithm is used to extract the line group that represent sea wave peak line, finally, the extracted wave peak line group image is mapped into HOUGH transform domain. Statistic and analysis are undergone then to obtain the precise wave length and direction. Because it takes a way similar to human vision to analysis nature texture image, this new method gives an expected good performance. Experiment result is also presented here.