基于区域生长的翡翠云雾纹理分割方法

Method of Emerald Cloud and Fog Texture Segmentation Based on Region Growing

根据翡翠云雾纹理特征,提出基于区域生长算法的翡翠纹理分割方法。首先,将图像转换到YCbCr颜色空间,根据统计的翡翠CbCr分量范围,利用边界跟踪算法确定分析区域,然后针对CbCr分量进行FCM聚类,选定生长区域种子,根据生长区域颜色相近和纹理相似的特征,进行区域生长,利用灰度共生矩阵分析Y分量进行纹理特征提取,针对分割效果影响因素,进行区域合并。实验表明该方法的分割效果良好。

this paper proposed a method of emeraId coud and fog texture segmentation based on region growing.First,convert the image to YCrCb coIor space, according to the statistics of the EmeraId CbCr component range,using the boundary tracking aIgorithm to determine the anaIysis area.Then FCM cIustering performed for CbCr components,so seed growing areas are seIected.Next,make the seeds growing for the regionaI characteristics of simiIar coIor and texture simiIar,and using the anaIysis GLCM texture feature of Y component extract factors for segmentation,and then merging simiIar region.