The Comparative Study on the Automatic Extraction Methods of Artificial Channel Based on ETM Image

It’s urgent for China to solve the water shortage. Quickly and accurately extracting water resources from satellite remote sensing has become an important means of the investigation and monitoring of water resources and wetland protection. The fact that the spatio-temporal span of channel is large made the investigation difficult especially by the conventional way. Remote Sensing plays an increasing important role in the water resources protection with advantages of large scale, integration, dynamics and fastness. The RS images recorded the truth of the surface landscape in history and can reflect the distributing and the status quo of the channel in different courses of history. The article analyses the spectral and spatial feature of channel in ETM images in order to extraction the channel automatically with the different RS methods combined with GIS technology. A comparison among these methods is made. In addition, the article assesses the results of single-band method and multi-band method qualitatively and quantitatively. This study provide a scientific basis for the protection of water resource.