Study on cloud extraction and snowfall estimation associated with cyclonic disturbances around the Antarctic using machine learning
| Project/Area Number |
16K21585
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Statistical science
Environmental dynamic analysis
|
|---|
| Research Institution | Hosei University (2019)
National Center of Neurology and Psychiatry (2018)
The Institute of Statistical Mathematics (2016-2017) |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2016-04-01 – 2020-03-31
|
| Project Status |
Completed (Fiscal Year 2019)
|
| Budget Amount *help |
¥3,380,000 (Direct Cost: ¥2,600,000、Indirect Cost: ¥780,000)
Fiscal Year 2018: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2017: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2016: ¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000)
|
|---|
| Keywords | 機械学習 / 降雪 / 南極 / 水蒸気輸送 / 衛星画像 / 衛星観測 / 雲画像 / 識別問題 / イメージング解析 / 画像解析 / ディープラーニング / 画像、文章、音声等認識 / 気象学 |
|---|
| Outline of Final Research Achievements |
It was confirmed that clouds with a high cloud top height, which are focused on during snowfall, have a contribution to snowfall at Antarctica and Syowa Station. Since the blizzard has a snow flying effect, it was judged appropriate that the cloud data extracted from the satellite cloud image as a parameter of the amount of the net surface snow mass balance.
I performed learning for automatic classification by CNN using cloud images of 5 years as a positive example of clouds during snowfall, but the accuracy is still insufficient, and further improvement of the accuracy of the learner is required. Based on these results, I did not practice on learning for a large image with rough pixels for 30 years.
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究は深層学習を長期間取得，保存されている南極域の衛星雲画像データに適用するものであり，データの利活用の一端を担っている．それと共にこれまで観測自体が難しい南極氷床質量収支を新たな方法で推測する手法の開発である．機械学習を南極域の衛星雲画像データに適用する例は世界的にはすでに存在しているが，降雪・涵養量の推定を目的とした研究としては初めてであり，気象・気候の数値モデルや衛星画像のアルゴリズムだけではない第三の手法として定量化が実現できればそれぞれのモデルの検証や改良にも役立つ結果となるだろう．
|
Report
(5 results)
Research Products
(16 results)
