Analysis of pixel overlaying proportion on cultivated land use change

As one of remote sensing change detection methods the post-classification method is often used to detect cultivated land use change. But this traditional approach does not fully consider the effects caused by mixed pixel in the process of marking the classified results. The article makes an improvement on this approach in that we think the result after classification is needed to make further division. We take advantage of the approach of texture analysis, and make further subdivision on the result of post-classification images. Finally we could reflect cultivated land use change in the form of overlaying proportion of pixels.