Video surveillance systems allow the monitoring of critical infrastructures to enhance their security. One of the main challenges of these systems lies in the need to involve personnel in surveillance for extended periods to ensure proper identification of potential risks. This task may be affected by worker fatigue or the quality of the captured image, factors that can negatively impact decision-making. Due to this, employing complementary systems that facilitate the detection of hazardous scenarios can be of great significance.
Researchers from the University of Castilla-La Mancha (UCLM) and the University of Sevilla (US) have collaborated on the development of DISARM, a system designed to identify scenarios involving weapons or violence using artificial intelligence. The purpose of this detection process is to promptly send real-time alerts to security agents, ensuring more accurate results and minimizing the occurrence of false positives.
More info available in the following website: https://projectdisarm.com
Weapon detection: The weapon identification by DISARM is based on object detection and recognizing the attacker’s pose when wielding a weapon. In the following link, there are several demonstrations: [link].
Fight detection: Through supervised learning, DISARM can identify conflict situations in real-time.
DISARM is a customized solution that can be modified and optimized by the research team upon request. New functionalities, such as the identification of bladed weapons, could be added to the system.
DISARM can be integrated into any existing surveillance camera, enhancing the monitoring activities and security of the following sectors:
- Government buildings
- Power plants
- Industrial complexes
- Sports stadiums
- Concert halls
- Educational institutions
The represented institution is looking for a collaboration that leads to commercial exploitation of the presented invention.
Financed: Proyectos de Prueba de Concepto (Agencia Estatal de Investigación)
Contact: Pablo Lago / email@example.com