Paper Abstract and Keywords |
Presentation |
2021-01-21 12:05
Examination of precipitation estimation using atmospheric variables Takanori Ito, Motoki Amagasaki, Kei Ishida, Masato Kiyama, Masahiro Iida (GSST Kumamoto University) NC2020-34 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In this paper, we developed a model for SR using ConvLSTM to improve the resolution of precipitation data.
In the related work, SISR using SRCNN, it is difficult to recover local values for precipitation data.
In this study, we propose a method that adds atmospheric variables to the low-resolution precipitation data and treats it as a time series data.
The proposed model is based on ConvLSTM, which treats images as time series.
In this evaluation, we compared the proposed model with the high-resolution precipitation data generated by SRCNN using the evaluation indices RMSE (Root Mean Square Error) and CC (Correlation Coefficient).
The results show that the proposed model is 0.93 times more accurate in terms of RMSE and 13.25 times more accurate in terms of correlation coefficient for high-resolution precipitation data. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Precipitation / Super Resolution / Convolutional Neural Network / Long Short Term Memory / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 120, no. 331, NC2020-34, pp. 13-17, Jan. 2021. |
Paper # |
NC2020-34 |
Date of Issue |
2021-01-14 (NC) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
NC2020-34 |
|