关键词: 土地资源类型, TM影像, 解译标志

Abstract: Using remote sensing and geographical information system digital processing techniques on advanced PC computer, we have established the land resources database of Northwest China through interpreting TM images (scale 1:100 000) in the middle period of 1990's. The studied region includes Shaanxi, Gansu, Ningxia and Qinghai provinces and its area is about 1 300 000 km². The land resources were classified into 6 classes and 29 sub classes: Ⅰ. Agriculture land: mountainous irrigated land, hilly irrigated land and level irrigated land; mountainous non irrigated land, hilly non irrigated land and level non irrigated land. Ⅱ. Forest land: woods, shrub, sparse woods and others. Ⅲ. Grassland: high coverage, moderate coverage, low coverage. Ⅳ. Water region: river and canal, lake, reservoir, glacier and for ever snow land, sea beach, flood land. Ⅴ. Residential and constructional areas: city and town land, rural residential land, industry and communication using land. Ⅵ. Non used land: sandy land, gobi, saline alkali land, swamp, bare soil land, bare rocky and gravel land, frigid desert, tundra land and others. In order to interpret correctly, we had worked out the interpreting key system of all sub classes land resources on TM images. Through analyzing relative reference and images we have corrected the interpreting key on the fields. The interpreting keys include: texture, appearance, size, distribution, geographical situation, color's hue and saturation etc.. Taking these keys as standards, we have completed the interpretation of land resources of Northwest China and finally had established a set of database that include images database, graphical database and attributing database. In the future, the interpreting key system that we have got can be widely used in the remote sensing investigation of land resources in Northwest China. Comparing with traditional interpreting, there are many advantages using digital processing method to interpret remote sensing images, such as, mapping rapidly, high accuracy and reducing expense. With the interpreting key, and graphical database and attributing database that well matched with images database, we can monitor accurately the land resources with the same classifying system and evaluating standards.

Key words: land resources type, TM image, interpreting key