基于二值化的数学形态学生物图像边缘检测算法

A edge detection algorithm of biologic Image based on dynamic binary and mathematical morphology

针对边缘检测算子对噪声敏感且检测边界相对模糊的缺点,提出了基于动态二值化的数学形态学边缘检测算法。该算法首先利用动态二值化方法确定将灰度图像转化为二值图像的阈值,然后利用数学形态学的腐蚀和去空洞相结合的方法去除二值化图像的噪声,最后用细化方法检测单像素图像边缘。仿真结果表明,该算法能降低噪声及图像模糊对边缘检测的影响,实现边缘的准确定位,并保留足够的图像细节,具有更强的去除噪声能力,为显微生物单体图像处理提供了一种有效的边缘细化检测方法。

A edge detection algorithm of biologic image based on dynamic and mathematical morphology, which can solve the affection of noise and blurring effect in image edge detection,is presented in this paper. The dynamic binary method was used firstly to define binary image threshold in transforming gray image to binary image. Secondly, the Erosion algorithm of mathematical morphology and de - empty method were used to eliminate noise of binary image. Finally the skeleton detection was used to detect image edge. The numerical experimental showsed that this algorithm reduced the effect of noise in image edge detection, and realized were accurately edge orientation, and hold adequately image detail. This algorithm is a good edge detection method in biologic image.