Keimyung University robotics students win top prize at maritime engineering CAE competition

A team of third-year robotics engineering students from Keimyung University took the top prize at the 2026 Maritime Engineering CAE Competition, the university announced Monday. Lee Hyung-jin, Kim Jae-yun and Son Yeo-won received a certificate and prize money of 1 million won at the awards ceremony held at Bexco in Busan on Thursday.

The team, advised by Professor Hong Sung-hoon, conducted research on a thermal imaging-based method for automatically detecting other vessels to ensure the safe navigation of autonomous ships in nighttime maritime environments.

The students focused on the difficulty autonomous ships face in detecting nearby vessels at night or in adverse weather conditions.

Standard RGB cameras lose detection performance in low-light conditions, while thermal imaging cameras are more resilient to changes in lighting but are limited by a shortage of maritime thermal imaging data needed to train AI systems.

To address this, the team developed a technology that converts existing maritime RGB footage into thermal imagery and uses it to train vessel-detection models.

They applied a content-preserving generative AI model that maintains the position and shape of ships, reducing the need for manual annotation of vessel locations and enabling the construction of large-scale training datasets.

The competition's final round, organized by the Korean Society of Ocean Engineers, drew 10 teams from institutions including Keimyung University, KAIST, Seoul National University, Hanyang University, Pusan National University and Korea Maritime and Ocean University.

The winning team also presented their research findings at the Korean Society of Ocean Engineers' autumn academic conference, held to mark the organization's 40th anniversary.

"It is meaningful that the students identified real technical challenges in the maritime field and proposed solutions using generative AI," Hong said. "We will continue to strengthen students' problem-solving capabilities through practice-oriented research that integrates robotics and AI."