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.