The Study on the Computer Automatic Classification of Land Cover in Chongqing Based on RS and GIS

The classification of Remote Sensing images is the very important step during the obtaining of the Remote Sensing information. There are two main methods in image classification: visual interpretation and computer automatic classification. The paper mainly discusses the second method. Taking Chongqing as an example, the information of the land cover type and the attributes of forests, grasslands, farmland, settlement, wetlands & water, and desert were extracted and obtained from the Remote Sensing images by computer automatic classification. The last conclusion shows that to the complicated area, such as Chongqing city, there were large errors in the automatic classification only depending on the computer, especially the type of deciduous needle-leaf forest and irrigated land, so plenty of corrections had to be done during the vectoring. Meantime, the combination of RS and GIS has inestimable advantages compared with the traditional methods. In the paper, the realizing steps and methods of how to classify automatically and how to process the data of land use by RS and GIS are provided.