Abstract: During the procedure of the stone image segmentation, there are some practical problems such as the great difference of stones and different regions of a piece of stone having distinctly different brightness characteristics. A growing method based on degree fussy number was introduced. During the region growing, different growing functions were selected. The conception of relative statistical character was put forward. Using this conception, not only can the seeds used for region growing be selected, but also can the size of stone and the main relative statistical character of the whole stone be known. All these above were the base of the region growing method based on different-size degree fussy number. Using this method, stone image segmentation can get a great result.

Key words: image segmentation, region growing, degree fussy number, relative statistical character

摘要： 在对石块图像的分割过程中,针对石块图像的形状差异大、同一石块的不同区域具有显著不同亮度特征的实际情况,采用了基于模糊隶属度的区域生长方法。在区域生长中,根据所得到的种子的不同特征,选用不同的生长函数。对种子的提取,提出了相对统计特征的概念。利用相对统计特征不仅提取了用于区域生长的种子,在此基础上,还初步知道了各个种子所对应的石块的大小和与整个石块相对应的初步的统计特征,这正是基于不同尺寸的模糊隶属度的区域生长方法的基础。通过这种方法,对石块图像的分割取得了很好的效果。

关键词: 图像分割, 区域生长, 模糊隶属度, 相对统计特征