The Application of Stationary Wavelet Transform in Crop Image Segmentation

摘要

According to the color and texture characters of the crop image, a method based on discrete stationary wavelet transform is presented to extract the crop image texture. Through two steps two dimension stationary wavelet transform, the crop image is decomposed into seven sub-images. From each decomposed wavelet field, the texture eigenvector can be extracted and combined with color character vector to form the image eigenvectors. By applying the fuzzy C-meaning cluster method, the image is segmented. The results show that the method can extract the crop image from the image with complex background.