Kookmin students win forest management AI contest

By Park Sae-jin Posted : August 24, 2026, 16:31 Updated : August 24, 2026, 16:31
Courtesy of Kookmin University

SEOUL, August 24 (AJP) - A team of students from Kookmin University has won the top prize at South Korea's inaugural forest artificial intelligence contest by developing a tool that helps private landowners predict 30 years of revenue and carbon absorption.

The web service, named MOFOM, addresses a longstanding challenge in South Korean forestry by democratizing access to fragmented government data. By simply entering a land parcel number, forest owners can instantly view economic and environmental projections across different management scenarios, turning dense academic models into a practical financial tool.

The students took the grand prize in the service creation category at the first Forest Science AI Use Contest, hosted by the National Institute of Forest Science. The competition drew 259 participants from 41 universities and 33 institutions to find technological solutions for regional and national forest management.

MOFOM integrates satellite imagery with government datasets, including national forest resource surveys, carbon absorption standards, and mountain weather records, to estimate timber volume and carbon content. It then generates six distinct management scenarios, ranging from immediate logging to long-term thinning and cultivation, calculating expected profits and carbon offset values for each path.

To ensure accuracy and reduce the risk of artificial intelligence hallucinations, the system relies on hard-coded calculations for its numerical data. The embedded language model is restricted strictly to explaining the finalized numbers, a feature that earned the team high marks for technical completeness and practical viability during live demonstrations.

Led by graduate student Choi Hee-do alongside undergraduates Na Jung-woo, Ha Su-beom, and Kim Min-seok, the team validated their model using actual land data and carbon offset project records from Boeun-gun, North Chungcheong Province. The students handled the entire development process, from satellite image analysis and economic calculations to web service deployment.

"It took a lot of time to connect forest data scattered across various institutions into a single service, but it is meaningful that we were evaluated for implementing it in a form that can actually be used," Choi said. "I want to expand the model verified in Boeun nationwide and develop it into a service that forest owners can actually use."

Copyright ⓒ Aju Press All rights reserved.