抄録
The Landsat satellite is one of the image data aquisition system for remote sensing purposes and has become the benchmark against other satellites. Thematic Mapper acquires data of 7 spectral bands and pixel size is 28.5 m×28.5 m at Tokyo. In order to analyze the Landsat Thematic Mapper Image in urban areas, it becomes important to discuss the conditions of mixels. Therefore, the authors made field investigation to obtain a ground-truth data concerned with and cover around the Imperial Palace. In the next place, the maximum likelihood classification method was applied to the Landsat thematic mapper image. Through the comparison of both information between obtained land cover mesh data (9.5 m×9.5 m) and classified image data (28.5 m×28.5 m), the conditions of mixels included in each category were analyzed. The following items become clarified. 1) Mixels are classified into some of categories involved in mixels themselves, in the case where categories are asphalt, shady asphalt, lawn, tree and water surface. 2) Mixels composed of asphalt and lawn, asphalt and water, almost belong to the category which the distribution ratio is higher in each mixel. 3) Mixels of asphalt and shady asphalt fall under the category of asphalt. 4) In disregards of the component ratio in mixels, mixels of tree and others scarcely belong to the category of tree. This leads to the underestimation of green fields.
