The maximum likelihood classification （MLC） is one of the most popular methods in remote sensing imag classification. Because the maximum likelihood classification is based on spectrum of objects

it cannot correctly distinguish objects that have same spectrum and cannot reach the accuracy requirement. In this paper

we take an area of Fengtai District of Beijing as an example and discuss the method of combining texture of high-resolution image with spectrum to improve the accuracy of TM image information extraction. Firstly

analysis of the textures of the in high-resolution imags is made by using texture analysis of Gray Level Coocurrence Matrices and selecting statistic index. Then threshold is selected and the optimal threshold is obtained according to entropy. Objects that have same spectrums such as vegetation land and cultivated land are distinguished using image segmentation in virtue of the optimal threshold. Finally

the find result is obtained through combining image segmentation with original classification. The finat result is compared with the classification results based on spectrum only or texture only. The result indicates that the objects with same spectrum are distinguished by using texture analysis in image classification

and the combination improves more than spectrum only or texture only in classification accuracy.