The Electronics and Telecommunications Research Institute (hereinafter referred to as ETRI) recently participated in the "PETS 2025 Challenge," an international competition on intelligent surveillance organized by the AVSS (Advanced Video and Signal-Based Surveillance) society, and demonstrated its world-class technological capabilities.
On September 3, ETRI announced that it had achieved first place overall in all tracking scenarios and second place overall in the detection category at the PETS 2025 Challenge.
(From left) Woojin Byun, Head of Daegu-Gyeongbuk Research Headquarters at ETRI; Sangwon Kim, Researcher; Gwangju Kim, Head of AI Infrastructure Research Division. Provided by Electronics and Telecommunications Research Institute
The PETS 2025 Challenge is an intelligent surveillance technology evaluation competition hosted by the University of Reading in the United Kingdom and sponsored by the European Maritime and Border Security Cooperation Project (EUMARS).
This competition comprehensively evaluates the ability to recognize and track people, vehicles, and vessels based on various sensor types and imaging conditions, including the three primary colors of light (RGB), thermal imaging, short-wave infrared (SWIR), and ultraviolet (UV).
It is characterized by presenting scenarios that reflect real-world surveillance environments, such as sensor resolution mismatches, changes in lighting and weather, and complex object congestion.
The PETS 2025 Challenge was conducted online. Participating teams analyzed the provided video datasets and submitted their results, which were then automatically evaluated by the organizers using internationally standardized metrics.
The evaluation results were officially announced at the AVSS academic conference. In the final rankings, ETRI secured first place in the tracking category and second place in the detection category, proving its top-tier global competitiveness.
ETRI implemented an artificial intelligence (AI)-based system that integrates the entire process of image recognition, detection, and tracking, maintaining high accuracy even in complex environments.
In particular, the system demonstrated stable performance and practical effectiveness even under challenging conditions frequently encountered in real surveillance systems, such as small objects, sensor resolution differences, and missing frames.
This achievement is the result of building upon and advancing ETRI's proprietary object tracking algorithms and the know-how accumulated over the past 30 years through national R&D projects in smart city intelligent traffic control and video analysis technologies.
ETRI expects this technology to be widely applied in fields such as urban traffic surveillance, port security, and maritime monitoring in the future.
In particular, it is expected to serve as a core foundational technology for next-generation smart city surveillance systems, as it operates reliably even under frequent changes in weather and lighting conditions.
Currently, ETRI is pursuing technology transfer discussions with domestic security and video solution companies, smart city platform operators, and port and maritime surveillance system manufacturers, aiming for commercialization by 2027. Once commercialized, the technology is expected to contribute to early detection of traffic accidents, optimization of traffic flow, monitoring of illegal vessels, and rapid response to maritime distress and pollution incidents.
Woojin Byun, Head of the Daegu-Gyeongbuk Research Headquarters at ETRI, stated, "This achievement globally demonstrates the excellence of Korea's tracking and detection technology, which operates reliably under various lighting, weather, distance, object size, and obstacle conditions. ETRI will continue to strengthen demonstration-focused R&D to contribute to public safety and the creation of social value."
Meanwhile, researchers Sangwon Kim, Dongyoung Kim, Kyungo Lee, and Gwangju Kim from the Daegu-Gyeongbuk Research Headquarters at ETRI participated in the PETS 2025 Challenge, overseeing technology development and system construction. In particular, postdoctoral researcher Dongyoung Kim drew attention from fellow researchers by presenting a digital twin-based, pixel-level early detection technology for fires and wildfires during an official AVSS session.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
