Research Topics

Geostatistics is a collection of statistical tools for analyzing spatio-temporal data. It has been widely used for spatial prediction or interpolation of sparsely sampled data. However, geostatisics can also be used for data integration, uncertainty modeling, and risk analysis as well as spatial prediction. Please keep in mind that geostatistics is NOT merely an advanced interpolation algorithm! We are trying to develop advanced geostatsitical algorithms and to apply them to environmental problems for supporting decision-makings.

• Integration of sparsely sampled hard data with auxiliary soft data and uncertainty modeling for environmental thematic mapping

• Value of information and risk analyses using stochastic simulation

• Cross-space-time-scale integration within a geostatistical framework

• Spatial analysis of disease data for medical geoinformatics

• Applications and Publications: Read more +

Remote sensing is the science of obtaining information through the analysis of data acquired from various man-made satellites and airplanes without making physical contact with the objects of interest. Our research focuses on environmental monitoring and thematic mapping with advanced methodological developments.

• Feature extraction and classification using statistical pattern recognition and machine learning algorithms

• Environmental monitoring, change detection, and thematic mapping

• Applications and Publications: Read more +

New advances in various data acquisition techniques have made it possible to rapidly collect large volumes of data sets. To make optimized decisions, better use must be made of all available information acquired from different sources or sensors. Data fusion or integration aims that 1 + 1 is more than 2! Multi-source data analysis, however, cannot be handled using the single source data processing and requires a new data analysis concept. We are developing efficient data integration models for various application fields.

• Development of data fusion models and systems (Bayesian probabilistic model, fuzzy logic, Dempster-Shafer theory of evidence, maximum entropy model, machine learning algorithms, etc)

• Applications and Publications: Read more +

Artificial intelligence (AI) techniques can extract undiscovered information from big data. However, AI models developed by the computer vision community must be modified to properly process geo-big data including satellite images, since the spatio-temporal nature of geo-big data should be accounted for during data processing. We are developing advanced AI models tailored for Earth observation and modeling with geo-big data. 

• Regression-based prediction, classification/change detection, and virtual image generation/super-resolution mapping using AI models including machine learning

• Applications and Publications: Read more +