Fousing on the agricultural lands controlled and managed by the Zhangtan Station of the Kaohsiung Irrigation Association, this study used FORMOSAT-II images and GIS system to differentiate and recognize the spatial and temporal usages of the agricultural lands through the pixel-by-pixel method as well as the zonal statistics method in ArcMap. With the use of the satellite images as well as their spectral responses, the agricultural land uses were classified into rice paddies, non-rice paddies, water bodies, and barren areas. The result shows that the use of NDVI can improve the accuracy of land-use classification from satellite images. The accuracy of land-use recognization by the ArcMap zonal statistics method is better than that by the pixel-by-pixel method. Effective and accurative classification of agricultural land uses is very helpful in effectively estimating the irrigation water requirements and the production of crops.
|Number of pages||15|
|Journal||Journal of Taiwan Agricultural Engineering|
|Publication status||Published - 2011 Jun 1|
All Science Journal Classification (ASJC) codes
- Agricultural and Biological Sciences(all)