The integration method of urban extension monitoring has been developed by using different period and multiplatform remotely sensed data in this paper. The better effect was obtained by using the integration method of classified images in the study of the rapidly developed Yueyang test area. In terms of urban information extraction of remote sensing image

various weight coefficient masks were designed to extract the relevant frequency vector as the urban contextual information. After progressively processing the urban image by using the contextual information

the final accuracy of classification is better than 91.7%. The test shows that the integration method of classified images is not only a flexible and practical method

but also a effective means to fully utilize the information sources of existed remotely sensed data.