CAJ | 학술논문

针对溢油 SAR 影像的油膜识别不准确的问题，该文引入最小模糊熵的图像分割方法，构建基于一维灰度直方图与二维灰度直方图的最小模糊熵隶属度函数，并基于模糊熵分割算法对试验影像进行处理。实验结果表明：最小模糊熵的分割算法能够对 SAR 影像实现比较好的油膜分割效果，且二维的模糊熵分割效果由于加入了图像灰度邻域信息，抗噪能力更强，分割效果要优于一维的模糊熵分割算法。同时，实验验证对于直方图双峰特性不明显的图像，采用最小模糊熵的的分割效果要优于最大模糊熵。
침대일유 SAR 영상적유막식별불준학적문제，해문인입최소모호적적도상분할방법，구건기우일유회도직방도여이유회도직방도적최소모호적대속도함수，병기우모호적분할산법대시험영상진행처리。실험결과표명：최소모호적적분할산법능구대 SAR 영상실현비교호적유막분할효과，차이유적모호적분할효과유우가입료도상회도린역신식，항조능력경강，분할효과요우우일유적모호적분할산법。동시，실험험증대우직방도쌍봉특성불명현적도상，채용최소모호적적적분할효과요우우최대모호적。
In order to identify the oil film on SAR image,this paper introduced the minimum fuzzy entropy image segmentation method,constructing the membership function based on the 1D-gray level histogram and 2D-gray level histogram respectively,and processed the experiment images based on fuzzy entropy segmentation method.The experimental result shows that the minimum fuzzy entropy image segmentation is more fitting for SAR images,and the 2D segmentation method is better than the 1D segmentation method,because it added the grayscale neighborhood information with strong resistance to noise. Meanwhile,the test shows that the image without two apex characteristic in histogram can get better result using the minimum fuzzy entropy than the maximum fuzzy entropy image segmentation.