基于数学形态学的边缘检测方法

针对图像的边缘检测问题，提出了一种基于顺序形态变换和局部熵相结合的边缘检测方法。该检测方法构造了基于百分位顺序形态变换的边缘检测算子，根据不同结构元素的尺度大小与结构元素的类型，选择不同的形态变换的百分值，从而使变换结果图像的局部熵最小，利用局部熵的特性，有效地锐化了图像边缘。实验结果表明，与传统的边缘检测方法相比，该方法具有很强的抑制噪声能力和良好的边缘检测性能。

Edge detection method based on math morphological transform

A algorithm based on percentile morphological transform and local entropy was proposed for edge detecting. This method selecting percentile morphological transform values were decided by structure size and structure type which were used on percentile morphological transform, in order to minimize the local entropy of local area. Experiment results show that, compared with the traditional edge detecting method, this method enhances the edge and decreases the noise.