Cotton leaf image edge detection using Mean-shift algorithm and lifting wavelet transform

Based on Mean-shift algorithm and lifting wavelet transform, a novel edge detection algorithm was proposed in this paper, in order to detect the edge of cotton leaf in an image with the presence of clutter and occlusion. Firstly, the color image was smoothed using Mean-shift algorithm. Then the lifting wavelet transform was used to enhance the edge of the smoothed image. Based on Canny operator, the edge of the cotton leaf was detected. The method can greatly reduce non-edge noises, and is able to effectively extract the edges between the overlapping leaves. Comparing with the experimental results of traditional edge detection methods, this approach can robustly detect the edge of the cotton leaf among clutter and occlusion while achieving obvious validity and accuracy.