Abstract: Traditional edge detection methods of ochotona curzoniae image have the problems of being sensitive to noise, large detail information of edge being filtered, having discontinuous and fuzzy detected edge. Aiming at these problems, the edge detection method based on mathematical morphology is used in the edge detection of ochotona curzoniae image in this paper. Using the corrosion and expansion algorithm of mathematical morphology and constructing proper structural element fitting to geometric properties of the ochotona curzoniae image, the edge of ochotona curzoniae image is detected. Experimental results show that the better effect of edge detection can be obtained by using this method.

Key words: ochotona curzoniae image, mathematics morphology, edge detection, structure element, feature extraction

摘要： 传统高原鼠兔图像边缘检测方法存在对噪声敏感、大量边缘细节被过滤、检测的边缘不连续、模糊等问题。为此，将基于数学形态学的边缘检测方法应用到高原鼠兔图像的边缘检测中。利用数学形态学的腐蚀和膨胀算法，构造适合高原鼠兔图像几何性质的结构元素，对图像进行检测。实验结果表明，该方法具有较好的边缘检测效果。

关键词: 高原鼠兔图像, 数学形态学, 边缘检测, 结构元素, 特征提取

CLC Number: 

• TN911.73