Multi-Scale Image Edge Detection Based on Wavelet Transform Modulus Maxima

This algorithm implements a multi-scale image edge detection method based on modulus maxima of wavelet transform. Wavelet transform serves as a mathematical tool for analyzing frequency components in signals and images. The algorithm employs wavelet decomposition across multiple scales to extract edge information from images. Specifically, we perform wavelet transformation on the image at different resolution scales, then identify edge locations by detecting modulus maxima points in the transformed coefficients. This multi-scale approach captures both coarse and fine edge features by analyzing image details across varying resolution levels, significantly improving edge detection accuracy and robustness. From an implementation perspective, the algorithm typically involves applying 2D discrete wavelet transform (using functions like wavedec2 in MATLAB), calculating gradient magnitudes and directions, then applying non-maximum suppression to locate modulus maxima points. The multi-scale framework allows combining edge information from different scales using fusion techniques to preserve significant edges while suppressing noise.