Abstract: An improved image edge detection method based on fuzzy theory and soft morphology is proposed.Firstly，an efficient membership function is chosen in ambiguity domain，and the fuzzy computational region is extruded by the concept of generalized fuzzy set，to increase the gray contrast of the edge.Then soft morphological nature and properties are used to choose a more efficient detector and structuring elements.Finally an ideal edge is got under the noise environment.Compared with other detectors this algorithm reserves more details of the edge with greater de-noising ability，and it has practicality and feasibility.

Key words: edge-detection, fuzzy, membership, soft morphology, detection algorithm

摘要： 提出了一种改进的基于模糊理论的柔性形态学边缘检测方法。在模糊域里选用效率更高的隶属度函数，借鉴广义模糊集的概念对模糊域计算区域进行拉伸和增强处理，增加边缘两侧灰度对比度，再利用柔性形态学特征及性质选取更有效率性能、更好的边缘检测算子和结构元，得到噪声存在条件下更加理想的图像边缘。与其他检测算法相比，该算法在取得更多边缘细节的同时具有更强的抗噪能力，具有一定的实用性和可行性。

关键词: 边缘检测, 模糊, 隶属度, 柔性形态学, 检测算子