Abstract: An algorithm based on wavelet lifting transform is proposed for improving multi-scale edge detection. It should be carried out the way the source image is decomposed by wavelet lifting transform so as to extract the edges of high-frequency and low-frequency sub-images separately. Then, the Canny operator is used for low-frequency information during edge detection, while the high-frequency component is detected after noise removal through multiplying the wavelet coefficients on adjacent scales. The final edge images are obtained through a fusion in accordance to certain rules. Experimental results showed that the algorithm is effective to extract image edges and has advantages of quick operation, efficient suppression of noise and high detection accuracy.