IREX Intrusion Video Analytics detects people and vehicles in sterile zones (e.g. secured perimeters, pipelines, critical infrastructure) using a deep learning (AI) algorithm with the extra-low rate of false alerts.
Unlike conventional intrusion detectors, IREX uses a convolution neural network (CNN) to detect and classify potential threats. Our CNN has been trained with hundreds of thousands of images and tested against the industry benchmarks such as the Imagery Library for Intelligent Detection Systems (i-LIDS) developed by a laboratory at the UK Home Office.
Additional, IREX Intrusion Video Analytics support:
- Pre-defined rules (bi-directional tripwires, polygonal regions of interest)
- Digital image stabilizer for more accurate tracking while the camera is shaking
- Custom event tags and camera tags (e.g. "north fence", "inner perimeter", "outer perimeter")
- Object filters by type (e.g. "person", "vehicle") with variable accuracy threshold
- Infrared lighting and thermal cameras