University of Aalborg and Milestone Systems collaborate on innovative project to reduce fatal accidents in Danish harbours by developing AI-trained models for thermal cameras to detect and respond to emergencies, setting new standards in safety technology.
In a collaborative project, the University of Aalborg and Milestone Systems have been recognised for significant strides in harbour safety technology. Utilizing artificial intelligence (AI) and video technology, the team developed a system aimed at reducing fatal accidents in Danish harbors.
The project, which focused on enhancing safety measures at Aalborg Harbor, earned the prestigious Danish Spar Nord Foundation Research Award. This accolade underscores the innovative approach taken by the research team, which concentrated on developing AI-trained models for thermal cameras to detect and respond to sudden, potentially life-threatening events such as individuals falling into the harbor.
Globally, drowning is a major concern, with approximately 236,000 people dying each year due to such incidents. Denmark has seen its share of water-related tragedies, with 1,647 lives lost between 2001 and 2015, and a significant portion of these occurring in harbor areas.
Recognising the urgent need for effective safety measures, researchers at the University of Aalborg employed AI combined with video technology to tackle this issue. Their work culminated in the creation of the largest outdoor thermal dataset for video analytics, covering an extensive nine-month period.
Replication of the specific activity of a person falling into the harbor poses ethical and practical challenges. To address this, the research team used a novel approach involving a test dummy filled with warm water, detectable by thermal cameras. This setup enabled the creation of an annotated thermal image dataset that could train AI models to detect fall incidents promptly.
The synthetic data greatly enhanced the models, introducing varied scenarios such as wheelchair users and cyclists, ensuring the AI system could handle diverse situations. Neelu Madan, a lead researcher and Ph.D. student at the University of Aalborg, highlighted the capability of the model to detect deviations from normal behavior and promptly trigger alarms for rescue teams.
“This model detects abnormal events like someone falling into the dock and immediately alerts rescue personnel, even if the incident occurs without any witnesses,” Madan explained. Beyond its immediate application, the research exhibits the broader potential of synthetic data in training AI models for various difficult-to-replicate scenarios.
Furthermore, Madan suggested the technology’s broader usability, such as in manufacturing, where it could monitor and enhance production processes by detecting anomalies that could lead to costly errors.
The project has set new standards in the intersection of AI and video technology for harbor safety. The advancements made through this research are anticipated to save lives and could have far-reaching implications in safety protocols within the transportation sector as they gain adoption worldwide.
Rahul Yadav, Chief Technology Officer at Milestone Systems, expressed excitement over the project’s potential impact. “We are incredibly proud of Neelu and the groundbreaking progress made at Aalborg Harbor. This collaboration represents a significant step forward in video technology, providing the community with the largest annotated thermal dataset.”
Milestone Systems, a prominent provider of video technology software, is focused on enhancing security and operational efficiency through data-driven solutions. Founded in 1998 and part of the Canon Group, Milestone’s technologies are implemented in over 500,000 customer sites globally.
Overall, the successful research partnership between the University of Aalborg and Milestone Systems underscores the transformative power of AI and video technology in enhancing safety and security within harbor environments and potentially beyond.
