基于均值—中值—梯度共生矩阵模型的最大熵分割算法*

针对基于灰度—梯度共生矩阵模型的最大熵阈值分割算法抗噪声差的缺点，引入了均值—中值—梯度共生矩阵模型，并提出了基于该模型的最大熵阈值分割算法。为了有效地节省计算时间与存储空间，进而导出了该方法的快速递推公式。实验结果表明，该算法优于灰度—梯度模型分割方法，并能抑制高斯噪声、椒盐噪声以及其混合噪声对分割结果的影响，提高了分割的鲁棒性。

Maximum entropy thresholding algorithm based on mean-median-gradient cooccurrence matrix model

In order to overcome the shortcomings of maximum entropy thresholding algorithm based on gray level-gradient co-occurrence matrix model with poor antinoise performance, this paper introduced a mean-median-gradient co-occurrence matrix model. Based on this model, proposed a maximum entropy thresholding algorithm simultaneously. For the purpose of saving computing time and storage space, presented a fast recursive method in the end. Experimental results show that the algorithm is superior to gray level-gradient model segmentation approach, and can suppress Gaussian noise, impulse noise and their hybrid noise, improves the robustness of the segmentation effectively.