Sin-Gil Kim
Department of Information Science, University of Sciences, Unjong District, Pyongyang, Democratic People’s Republic of Korea

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

Time Series Prediction of Near-Surface Temperature Using an LSTM Neural Network Sin-Gil Kim; Un-Song Ryu
Indonesian Journal of Earth Sciences Vol. 5 No. 2 (2025): July-December
Publisher : MO.RI Publishing

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52562/injoes.2025.1482

Abstract

Predicting air and soil temperature is an active research topic in response to adverse changes in the global ecological environment, such as global warming and biodiversity loss. To improve prediction performance, various machine learning models and methods have been applied, some of which are costly to implement because they use spatiotemporal data such as global numerical weather prediction data. This study aims to establish a simple method for constructing a predictive model with relatively high accuracy. In this paper, we propose a method to construct a prediction model using a long short-term memory (LSTM) neural network with only time series data of measured temperatures at certain time intervals at stations in the region. The proposed method was applied to predict daily minimum and maximum temperatures for up to 10 days using only time series of near-surface temperatures observed at meteorological stations in Pyongyang and in coastal and inland areas of the Democratic People’s Republic of Korea during 2019–2023. The results show that the proposed model achieved relatively high and stable prediction performance using only observed data. The mean accuracy within ±2 °C (ACC) was 73.70% for 4-day and 87.97% for 10-day maximum temperature predictions.
Fusion and Super-Resolution of Remote Sensing Images Using Deep Learning and Gram-Schmidt Orthogonalization Sin-Gil Kim; Il-Kwang Jo; Gyong-Song Chae; Ji-Chon Ro
Indonesian Journal of Earth Sciences Vol. 6 No. 1 (2026): January-June
Publisher : MO.RI Publishing

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52562/injoes.2026.1632

Abstract

Multispectral satellite images contain rich spectral information but often have low spatial resolution. Super-resolution of these images is important for resource development, environmental protection, disaster prevention, regional ecology, and change detection. Deep learning models, including convolutional neural networks (CNNs), have been widely applied to the super-resolution of satellite images, including Sentinel-2 images. However, their output cannot exceed the maximum spatial resolution of the sensor. Here, we propose a method that first enhances the 20 m Sentinel-2 bands to 10 m using a CNN and then fuses them with 1 m aerial images by Gram-Schmidt orthogonal transformation and its inverse, while preserving multispectral information. The method was tested on the fusion of aerial and Sentinel-2 images to visualize vegetation distribution in urban areas. The proposed network achieved a root mean square error (RMSE) of 43.5 and a correlation coefficient (CC) of 0.941, compared with 123.4 and 0.811 for bicubic interpolation and 117.8 and 0.866 for area-to-point regression kriging (ATPRK). The fused images appeared as sharp as the aerial images while preserving the shape of the original spectral curves.