角点检测是计算机视觉处理的首要步骤 ,本文提出一种平面曲线角点检测的方法。首先 ,从人类视觉感知出发 ,给出角点两个重要性质作为对传统角点性质的补充 ,基于上述两个性质 ,模糊集合的概念被引入到检测问题。然后 ,给出三组包含角点隶属度的特征提取公式 ,综合三组特征 ,给出角点检测、定位、优选的判据。文中最后给出算例检测结果和感兴趣部分的特征曲线 ,以及对历史文献测试图像的检测结果。结果表明 ,本文算法使用模糊集合理论 ,在实现上非常简单 ,检测效果也很理想 Corner detection is the first step of computer vision processing. This paper presents a method of detecting the corner point of planar curve. First of all, starting from human visual perception, two important properties of corner are given as complements to the nature of traditional corner. Based on the above two properties, the concept of fuzzy set is introduced into the detection problem. Then, three sets of feature extraction formulas including corner membership degree are given. Based on the three sets of features, the criteria of corner detection, location and optimization are given. In the end, the test results of the example and the characteristic curves of the part of interest are given, and the test results of the historical document test images are given. The results show that the proposed algorithm uses fuzzy set theory and is very simple to implement and the detection result is very satisfactory