海上溢油SAR图像分割算法研究

Study on segmentation of SAR image for oil spilled at sea

运用SAR进行海上溢油监测是目前主要的手段。本文采用ENVISAR-ASAR数据,在分析溢油特性基础上,基于专家知识,开展一种区域种子扩散的方法进行图像分割研究,为溢油信息自动提取打下基础。该方法能够有效的分割出溢油信息,采用不同的预值,产生不同的结果,这与专家知识有关,这在一定程度上使分割结果受到人为影响,并且在分割后的图像的平滑与形状上存在误差,这将在以后的研究中进一步研究。

The SAR is a useful way for the monitor oil spilled at sea. A zone diffusion to segment SAR image was developed based on the analysis of the oil spilled characters and expert knowledge, paving for oil extracting. This method can give the effective segment SAR image, however, the different primary value given might obtain the different outcomes, which is relation with the expert knowledge and impact the segment results in same extent, also image smooth and pattern will be further studied.