Classification Of The Remote Image

The classification of remote sensing image is a process that all the pixels in the image are seperated into some species by their characters. The general ways are Surpervised classification and Un-supervised classification. Un-surperwised classification is a process of clustering, Supervised classification is a process of learning and training, it need enough pre-knowledge. Without enough pre-knowledge, the un-supervised classification is used in this paper. In the paper, the traditional un-supervised classification such as K-means,Isodata are shown detailedly. Based on this ,K-means in the Gauss distribution , Rayleigh distribution and fuzzy Isodata are introduced. From the results ,we can see that the classified effect is much better than the traditional ways. In the data analysis ,the Hidden Markov Model in Wavelet domain used in SAR classification and its effect are shown too. So we can draw a conclusion that the new ways in classification in the paper which have been shown are More accurate and flowing than the traditional ones. Aiming at the specie noise of the SAR image ,we can see that the HMM is an effective method for the noise.