2.3. Prescription of multi-temporal satellite data
2.3.1. Vegetation index
The Enhanced Vegetation Index (EVI) can be linearly correlated with the leaf area index, and has a higher sensitivity than that of NDVI in high biomass areas (Huete et al., 2002). Since the biomass of paddy fields is low compared with forests, EVI retains linearity in paddy fields. Therefore EVI can be used to evaluate the process of paddy growth adequately with a wide dynamic range. Furthermore, EVI is more practical than NDVI when humidity is high. Since the rice growing stage falls in the rainy season when humidity is high, EVI is a more practical than NDVI for monitoring the paddy fields in Japan. Therefore, we used EVI in our method.
2.3.2. Removing data affected by thick clouds, and sensor zenith angle