Methodological Approach for the Detection of Ground Cover Classification and Surface Temperatures of the Waterfront using Landsat TM Data

摘要

In this research, natural areas and urbanized areas in a waterfront of southern Osaka were classified according to ground cover using Landsat TM data, and an attempt was made to improve the accuracy of such classifications. DN values from band 6 were then used to investigate the relationship between the above ground cover classifications and surface temperature. It was found that accuracy was high in classifying forested and other green categories. In urbanized areas, which have long been thought difficult to classify accurately, it was found that uniform ground cover conditions generate uniform band characteristics. Thus it was demonstrated that even urbanized areas can be classified with some degree of accuracy. Regardless of classification category, areas where the ground cover was urbanized areas had a higher surface temperature than other areas, and it was found that the sea had the effect of lowering surface temperatures on adjacent land.