Integrated Spectral And Spatial Information Mining In High-resolution Remote Sensing Imagery

How to abstract information from remote sensing image effectively is the key problem in remote sensing image processing. The conventional approach abstracting information only based upon the pixel level, which take no account of the spacial correlativity between the neighborhood pixels. Especially with the improving of the high spatial resolution remotely sensed images, the traditional spectral-based methods have proven inadequate for the classification.In the background, the dissertation is based on the remote sensing image interpreter principal and wavelet MRA theory .Using the strongly matrix operation capability of Matlab software to mine the image information and do the quantitative application. Integrated spatial and spectral information to improving the classification accuracy. We apply the algorithm to the typical cases of IKONOS, and compared the experience result with the conventional approach. we do some primarily research on the mixed classifier integrated spectral and spatial feature.