基于图像纹理自适应的边缘检测算法

Edge Detection Algorithm Based on Image Texture Adaptively

通过对图像进行Fourier分析,获取能够表达图像纹理的能量谱,进而定义能量谱的环特征和楔特征。环特征可以描述纹理的粗细,楔特征可以描述纹理的方向,以这两个特征值作为边缘检测模板选择的依据,使边缘检测具有纹理自适应能力。实验表明,该算法边缘检测精确,抗干扰能力强。

Through the Fourier analysis of input image, we can obtain its power spectrum which implies image texture. Further the ring and the wedge character of spectrum can be defined. The ring character can describe the density of image texture. The wedge character can describe the direction of image texture. Therefore, the algorithm proposed can detect image edges based on texture adaptively, by using these two characters as criteria of template selection. The experiments show this algorithm can detect image edge accurately and has the capability of noise-resistant.