The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B3b. Beijing 2008 
unurban 
1.8 
98.2 
Figure 6. Urbanized area extracted using skewness 
4.2 Classification Result and Accuracy Assessment 
Follow the steps in 4.1, based on ETM+ images of Beijing 
Metropolitan area and DEM data, the urbanized area is 
extracted. Through visual interpretation, the profde of extracted 
urban area highly accorded with the real situation. 
To compare classification accuracy of different approaches, 
taking visual interpretation as the true value, 40 patches(22572 
pixels) of urbanized area and 40 unurbanized patches (45011 
pixels) were selected to generate the confusion matrix: 
Table 1. Comparison of accuracy with different classification 
approaches 
Table 1 shows skewness can effectively improve the 
classification accuracy compared with traditional approaches. 
So the use of skewness provides a new way to extract the 
urbanized area accurately. 
5. CONCLUSION 
As the fast urbanization occurs in China, setting up a large 
extent observing system on cities is an urgent task for Chinese 
government. And ETM+ images from 1999 to 2004 provide an 
abundant and cheap data resource for such task. Combined with 
multi-spectral bands, the skewness approach proposed from this 
paper fully exploits the detail information from band 8, so this 
is a new way for extensively observing the urban pattern and its 
change. 
REFERENCES 
Armenakis, C., 2003. A comparative analysis of scanned maps 
and imagery for mapping applications. ISPRS Journal of 
Photogrammetry and Remote Sensing, 57(5), pp.304-314. 
Chiang, S., 2001. Unsupervised target detection in hyperspectral 
images using projection pursuit. IEEE Transactions 
Geoscience and Remote Sensing, 39(7), pp. 1380-1391. 
Hlavka, C., 1981. Land-use mapping using edge density texture 
measure on Thematic Mapper Simulator data. IEEE Transaction 
On Geoscience and remote sensing, 1, pp. 104-108. 
Irons, J., 1981. Texture transforms .of remote sensing data. 
Remote Sensing of Environment, 11, pp.359-370. 
ACKNOWLEDGEMENTS 
This research was supported by the Wiser Foundation of IDC - 
Peking University (NO. W08SD04 ).