一种基于数据融合和小波变换的图像边缘检测方法

论文提出利用数据融合和小波变换进行图像边缘检测的一种方法。此方法首先对同一地区的多谱段图像用小波变换进行融合预处理，然后直接采用小波变换系数动态地调整边缘判别的阈值，对融合图像进行边缘检测。试验结果表明，此方法不仅能有效地抑制噪声，而且对具有多种边缘特征的图像均有良好的适应性。

A Method of Edge Detection Based on Wavelet Transform and Data Fusion for Multiple Images

Edges of an image reflect the information of the im age mostly. They co ntain the basic characteristics of the image. The traditional edge detection met hods c an be divided into two categories, time field and frequency field. In time field , there are Sobel edge operator, Prewitt edge operator, Kirsch edge operator, th resholding, etc. In frequency field there are many kinds of filters of differen t bands widths. Methods that only enhance the edge like Sobel edge operator can not fit as some parts of the image are flat and some parts are acute . In frequency field, normal Fourier transform has no resolution in time field. Any modification of any frequency will influence the whole time zone leading to bad behaviour. Even though the short timed Fourier transform's effect is limited , wavelet transform is employed as a new method of information proc essing. Wavelet transform can focus on any detail of the analyzed object by ta king more and more fine step of time field or space field. It is better adapted to the acutely varied edges than normal Fourier transform. This paper provides a method that detects the edge of multi_bands fused images based on wavelet t ransform. With this method the multi_bands images of the same area are fused wit h the wavelet transform. Then the image is reconstructed and the edge of the fused image is detected. The coefficients of wavelets are applied directly to adjust the threshold dynamically in the edge detection. The result of the exp eriment indicates that this method not only restrains noise effectively,but also adapts to the images with more edge characteristics.