基于均值—梯度共生矩阵模型的最大熵多阈值算法

Multi-threshold Algorithm Based on Maximum Entropy Of the Mean Value-Gradient Co-occurrence Matrix

为解决噪声显微细胞图像的多阈值分割问题,该文提出基于均值和梯度共生矩阵模型的最大熵多阈值算法。选用象素点的邻域灰度均值和梯度值构成二维灰度直方图。因为对象素点取均值可以平滑噪声,取梯度值可以锐化边缘,所以该算法能够改善图像的分割质量。考虑显微细胞图像多阈值分割的要求,该算法对二维灰度直方图采用改进的区域划分方式。通过优化传统的求熵算法,来减少运算时间,使之更加适合于擅长矩阵运算的MATLAB编程语言,从而提高运算速度。实验证明,该算法去除了噪声干扰,实现了显微细胞图像的多阈值分割,运算速度较快。

In the paper,the multi-threshold segmenting algorithm of a polluted micro cell image was proposed.Based on the maximum entropy of the mean value-gradient co-occurrence matrix of images the algorithm performs the multi-threshold segmentation of the cell images polluted by the noises.The mean value of images can smooth noises well and the gradient can preserve more information of the image edge.Therefore,combination of two approaches will improve the segmentation result.In the paper the traditional division method of the entropy calculating the region was changed in order to perform multi-threshold segmentation.The traditional entropy algorithm was improved for programming of the MATLAB language and for getting the thresholds more quickly.Experimental results show that the algorithm can increase the speed of operation,raise the power of resisting noises and improve the segmentation effects.