基于S变换的图像边缘检测算法研究

针对边缘检测中存在的噪声敏感性问题，提出了一种基于S变换的改进型边缘检测算法。S变换是一种时频分析工具，它能够很好地检测信号的奇异性和非正则性，而在图像中，图像灰度的奇异性代表着图像的边缘。通过将S变换后相邻频率的系数进行乘积可以有效地抑制噪声，增强边缘，采用一个阈值计算方法对相乘后的S变换系数进行阈值滤波，整合来自两个正交方向的边缘得到最终的图像边缘。实验结果证实了该算法的有效性。

Research of image edge detection based on S-Transform

To reduce the effect of the noise on edge detection, an improved algorithm of edge detection is proposed based on S-Transform. The S-Transform（ST） is a time-frequency analysis tools and has ability to detect the singularities and irregular structures in signals which often carry the most important information in signals. In images, the singularities of the intensity indicate the edge. Multiplying the S-Transform coefficients in the adjacent frequencies can depress the noise and enhance the edges effectively, and then a method is proposed to calculate the proper threshold to find the real edge in the coefficients. Finally, a scheme is designed to synthesis the two edge maps obtained in two orthometric directions. A set of the experiments demonstrate the effective of the approach.