Better Data for Better AI
Delivering Better Accuracy and Lower False Alarms
What if LIDAR could cost-effectively be integrated into surveillance cameras?
What if 3D data could be used to enhance existing surveillance AI and detection algorithms?
Oyla’s 3D-aware surveillance camera resolves previously intractable issues of variable lighting, occlusion, and the true distance and scale of objects. Oyla does this by using its inherently spatially fused 3D “Depth” and video streams to produce enhanced RGB video (e-RGB). e-RGB video can be utilized by existing AI and detection algorithms to achieve greatly increased performance. In addition, Oyla uses the 3D data to accurately detect intrusions into user-defined perimeters.
AI analytics are increasingly applied to surveillance video streams to automatically flag unusual activity in real time, but their accuracy falls short. Oyla’s approach is fundamentally different: obtain better data as input to the AI in the first place. To this end, Oyla has developed a unique AI-based video surveillance system which augments the video with a an inexpensive proprietary 3D sensor. The 3D sensor is integrated with the video sensor at a hardware level, and collects 3D data across the video field of view at up to 50 meter range. The 3D data stream is independent of the video stream, and so statistically improves the accuracy of the video AI. This 3D sensor provides its own infra-red illumination and so can operate in complete darkness. It can also be used to define very accurate intrusion detection perimeters.