基于最大模糊总熵准则自动选择灰度图像分割的最优阈值

图像不仅含有由统计不确定性产生的信息量 ,而且含有模糊不确定性产生的信息量。我们通过用来测度统计信息的香农熵为模糊总熵 ,开发出基于最大模糊总熵准则的最优阈值技术 ,它是熵阈值技术的良性拓广 ,实验表明它能在二值化后保留更多的图像信息 ,取得更好的图像分割效果。

Threshold Selection Based on Maximal Total Fuzzy Entropy Criterion

In a gray level image there exists the information contained in fuzzy non determinacy as well as the information contained in statistical non determinacy. By means of extending in a well posed Shannon entropy that measures statistical information to total fuzzy entropy, we develop an optimal threshold selection algorithm based on the maximal total fuzzy entropy criterion. The algorithm is the well posed extension of the threshold selection algorithm based on a statistical entropy criterion. The experiments demonstrate that the technique can be able to keep more details in the thresholded image and have a better image segmentation effects.