When classifying the hyper-spectral remote sensing images using the max-likelihood classifier, the determinant of spectra s cov-matrix is singular approximately and an unreasonable result comes out because of the big number of the bands and the close correlation between them.

Then based on the enhancement of extraction with maximum likelihood classification and compared the result with the effect that used maximum likelihood classification only.

In this paper,RRI(Ratio Resident-area Index) and Maximum Likelihood Classification(MLC) were used to retrieve urban residential areas in the region of Xi an,respectively,from the satellite image of Landsat TM in 2003.

This paper introduces the characteristics of and the existing problems in the maximum-likelihood classification,advances some methods that should be adopted in solving these problems,and discusses on the classification technology of artificial neural network and the image texture features participating classification.

Algorithm research of maximum likelihood classification based on fuzzy inference

Considering two stages probability method of maximum likelihood classification(MLC),this article proposed a new method of exploiting spatial information to improve classification rules by adjusting the prior probability according to the local spatial information.