Abstract: The image edge detection is an important part of the image processing.For the image edge and noise are with the high-frequency characteristics and the distribution of noise is always unknown，the single-scale edge detection is hard to detect the image edge accurately.However，the 2-D A Tuous algorithm based on wavelet can detect the image edge in different scales.It detects the details of the edge in the small-scale，and detects the contour of the edge in the large-scale.This paper introduces the basic principle and operator implementation of the 2-D A Tuous algorithm，and compares it with traditional methods through the experiment.The experiment shows that this algorithm can detect the image edge effectively，and it is better than the traditional methods.

Key words: A Tuous algorithm, multi-scale analysis, edge detection

摘要： 边缘检测是图像处理的重要内容。由于图像边缘和噪声都具有高频特性，而且噪声的分布往往是未知的，所以用单一尺度的边缘检测很难准确地检测图像边缘。而基于小波的二维A Tuous算法可以检测不同尺度下的图像边缘。在小尺度下检测边缘细节，大尺度下检测边缘轮廓。该文介绍了二维A Tuous算法的基本原理和算法实现，并通过实例与传统方法比较。实验结果表明，该方法可以有效地检测图像的边缘，并优于传统方法。

关键词: A Tuous算法, 多尺度分析, 边缘检测