| Budget Amount *help |
¥6,240,000 (Direct Cost: ¥4,800,000、Indirect Cost: ¥1,440,000)
Fiscal Year 2019: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2018: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2017: ¥3,510,000 (Direct Cost: ¥2,700,000、Indirect Cost: ¥810,000)
|
|---|
| Outline of Final Research Achievements |
This study investigated the applicability of convolutional neural network (CNN) and convolutional encoder-decoder network (ConvED) to develop a new concept of rainfall forecasting model. CNN and ConvED are well known machine learning algorithm that is specialized in image recognition.
In this study, three-dimensional spatiotemporal data was created with the time series of multiple atmospheric variables from Amedas point gauged data and Himawari satellite observation data. This three-dimensional data array (time-space-variable) is treated as an image with multiple color channels, and it is utilized into CNN and ConvED algorithms to predict rainfall occurrence and rainfall amount in 30 min lead-time.
|
