基于多尺度小波变换的边缘检测算法

提出一种用多尺度小波变换进行边缘检测的算法,并将该算法与经典的Sobel算子进行比较,结果表明用该算法进行边缘检测是可行的.在多尺度信号边缘检测中,考虑到信号的边缘不仅仅定义为信号奇异性的表现,而且也是视觉的一种反映,它与人的视觉特点,先验知识紧密相关.而信号的孤立奇异指数可以由小波变换在该点随尺度参数变小时的衰减速度确定.由于小波变换的上述特征,采用小波变换对图像进行边缘检测非常有效.实验结果表明,本方法和传统的边缘检测算法相比具有定位精度高,去噪效果好等优点.

The Algorithm of Edge Detection Based on Multiscale Wavelet Transform

This paper proposes a multiscale wavelet transform approach for edge detection.Comparing this algorithm with Sobel algorithm,the experiment result shows that this algorithm is available.In the multiscale signal edge detection,considering that signal edge detection is not only the representation of signal singularity but also a visual reflection,it is closely bound up with visual characteristic of human beings and a priori knowledge.The signal isolated singularity index can be determined by the rate of decay of wavelet transform with the diminishment of scale-parameter at the same point.Therefore,in virtue of its above mentioned characteristics,it is very effective to adopt wavelet transform for image edge detection.