We turn images of the Earth into spatial models that machines can reason about and act on. Our work starts with satellites and photogrammetry and now extends to drones, geospatial foundation models, LLM agents, and physical AI for industry.
Our research follows the path spatial data takes: sensed from above, rebuilt as geometry, understood by learning models, and used by machines that act in the real world. Satellite remote sensing and GeoAI for the atmosphere, ocean, cryosphere, and cities are our base. As part of an AI engineering department, we are pushing each stage toward engineering systems that industry can deploy.
Capture the world from orbit, air, and ground.
Turn gappy observations into complete fields and geometry.
Learn what the spectra and geometry mean.
Close the loop in the physical world.
Within an hour of campus there are tidal wetlands, one of the world's largest steelworks, a national petrochemical complex, and wide farmland. Few regions offer this mix of natural and industrial sites to test spatial AI on real problems.
Extend our satellite salinity and SST reconstruction to the bay, and map reed beds and tidal flats with drones to estimate the carbon they store.
Build 3D models of plants, ports, and yards, and use them to plan inspections, track stockpiles, and train robots in simulation before deployment.
Detect thermal anomalies, leaks, and surface deformation around industrial complexes using thermal drones and InSAR.
Monitor crop health, yield, and water stress from multispectral imagery, with foundation models that transfer across seasons and crops.
Deep learning for satellite and radar data across four Earth systems. D. Han is in bold; the full list of 41 works is on Google Scholar.
Exploring the potential of machine learning post-processing to generate ERA5-consistent atmospheric profiles from geostationary satellite retrievals
D. Han, M. Choo, S. Jung, J. Lee, H. Choi, J. Im
Remote Sensing 18(14), 2310
CARE-SST: Context-aware reconstruction diffusion model for sea surface temperature
M. Choo, S. Jung, J. Im, D. Han
ISPRS Journal of Photogrammetry and Remote Sensing 220, 454–472
PARAN: A novel physics-assisted reconstruction adversarial network using geostationary satellite data to reconstruct hourly sea surface temperatures
S. Jung, J. Im, D. Han
Remote Sensing of Environment 323, 114749
Exploring the potential of latent embeddings for sea ice characterization using ICESat-2 data
D. Han, M. Karimzadeh
IGARSS 2025
Long-term prediction of Arctic sea ice concentrations using deep learning: Effects of surface temperature, radiation, and wind conditions
Y.J. Kim, H. Kim, D. Han, J. Stroeve, J. Im
Remote Sensing of Environment 318, 114568
Deep learning-based gap filling for near real-time seamless daily global sea surface salinity using satellite observations
E. Jang, D. Han, J. Im, T. Sung, Y.J. Kim
International Journal of Applied Earth Observation and Geoinformation 132
Precipitation nowcasting using ground radar data and simpler yet better video prediction deep learning
D. Han, M. Choo, J. Im, Y. Shin, J. Lee, S. Jung
GIScience & Remote Sensing 60(1), 2203363
Key factors for quantitative precipitation nowcasting using ground weather radar data based on deep learning
D. Han, J. Im, Y. Shin, J. Lee
Geoscientific Model Development 16, 5895–5914
Remote sensing of sea surface salinity: Challenges and research directions
Y.J. Kim, D. Han, E. Jang, J. Im, T. Sung
GIScience & Remote Sensing 60(1), 2166377
Prediction of monthly Arctic sea ice concentrations using satellite and reanalysis data based on convolutional neural networks
Y.J. Kim, H.C. Kim, D. Han, S. Lee, J. Im
The Cryosphere 14(3), 1083–1104
Improving local climate zone classification using incomplete building data and Sentinel-2 images based on convolutional neural networks
C. Yoo, Y. Lee, D. Cho, J. Im, D. Han
Remote Sensing 12(21), 3552
A novel framework of detecting convective initiation combining automated sampling, machine learning, and repeated model tuning from geostationary satellite data
D. Han, J. Lee, J. Im, S. Sim, S. Lee, H. Han
Remote Sensing 11(12), 1454
Comparison between convolutional neural networks and random forest for local climate zone classification in mega urban areas using Landsat images
C. Yoo, D. Han, J. Im, B. Bechtel
ISPRS Journal of Photogrammetry and Remote Sensing 157, 155–170
Convolutional neural network-based land cover classification using 2-D spectral reflectance curve graphs with multitemporal satellite imagery
M. Kim, J. Lee, D. Han, M. Shin, J. Im, J. Lee, L.J. Quackenbush, Z. Gu
IEEE JSTARS
Published in RSE · ISPRS J · The Cryosphere · GMD · GIScience & RS · JAG · IEEE JSTARS. Next: CVPR · ICRA · NeurIPS.
All 41 works on Google Scholar →PRISM is young, so the first members shape its direction, tools, and culture.
Satellite remote sensing and deep learning for the atmosphere, ocean, and cryosphere. Dept. of Artificial Intelligence Engineering, SCNU.
Geospatial foundation models, spatiotemporal forecasting, or physical AI.
Drone photogrammetry, 3D reconstruction, GeoAI.
Any year. Curiosity and Python are enough to start.
PRISM 연구실에서 학부연구생 및 석·박사과정 학생을 모집합니다.
You will work on real data from satellites, drones, and industrial sites, and ship models that run outside the notebook. Students from AI, computer science, geoinformatics, civil, mechanical, and environmental engineering are all welcome.
Join a project, learn the tools, and co-author a conference paper.
Lead your own research line from data collection to publication.
Bring a spatial problem from your site. We co-design data, models, and field trials.