Abstract
Remotely sensed data is expected to predict grain yield rapidly and extensively. So some researches have been carried out. But the problem is that the difference of growth stage makes the difference of spectral characteristics.
Therefore, growth index is proposed to correct the difference of growth stage.
Using Landsat TM data and the growth index of paddy fields in Chiba prefecture, rice yield per unit area in each pixel is estimated with regression models.
The rice yield is estimated more precisely from the growth index than from a simple digital number set. Furthermore, it is suggested that the information of variety is important for rice yield prediction from remotely sensed data.
