边缘是图象的特征,对许多应用领域的图象分析及分类是很有用的。大量的应用和对边缘定义及特征的探讨,促进了大量边缘检出算法(或算子)的研究。一些算法在特定的应用中可能是有效的,但对另一些应用则不然。本文主要描述各种不同的边缘检出算法和根据已知的性能判定方法,评价性能。这一评价是以在计算机上对不同形状、斜率和背景噪声的边缘仿真为基础的。与其它有关文章的结论结合在一起,有助于鉴别不同的边缘检出方案,更好地了解所使用的方法的效果和局限性。 Edges are image features that are useful for image analysis and classification in many applications. The large number of applications and the discussion of edge definition and features have promoted a large number of edge detection algorithms (or operators). Some algorithms may be effective in a particular application, but not others. This article describes a variety of different edge detection algorithms and performance evaluation methods based on known performance evaluation. This evaluation is based on edge simulation of different shapes, slopes and background noise on the computer. Combined with the conclusions of other related articles, it helps identify different edge detection scenarios and better understand the effects and limitations of the methods used.