VASTIQ Solutions, under the leadership of Ts. Nicholas Morgan, is at the forefront of AI-driven innovation with the development of VAST-NXT.ai, an advanced AI video analytics solution. This cutting-edge system is designed to provide accurate, real-time data analytics for businesses, integrating RTSP URL streams into a centralized dashboard for seamless management. By enabling customers to monitor multiple video streams efficiently, VAST-NXT.ai aims to enhance situational awareness and deliver precise, data-driven reports. The initial phase of development is focused on ensuring smooth RTSP integration, allowing users to track key metrics effectively
For Phase 2, the R&D team is expanding VAST-NXT.ai’s capabilities with three major enhancements: Tracking ferry operations by allowing users to manually input schedules and departures, with future automation possibilities under discussion; developing a signal-based alert system to indicate when a predefined number of people have been counted; and Improving the AI model’s ability to accurately distinguish between children and adults, refining classification accuracy through advanced deep learning techniques. These upgrades will significantly improve the efficiency of AI-based video monitoring, particularly for transportation and crowd management applications.
A critical component of this development is hardware optimization, particularly GPU performance for AI processing. Based on VASTIQ's internal benchmarking, an RTX 3050 (local system) can process 10 camera feeds with a low-accuracy model, and 3-5 feeds with a high-accuracy model.
Scaling up, the RTX 4060 Ti can support 40 low-accuracy or 20 high-accuracy streams, while the RTX 4070 Ti Super can handle 80 low-accuracy or 50 high-accuracy streams. These insights are vital for ensuring that the AI model operates at optimal efficiency without hardware bottlenecks.
Additionally, the R&D team is prototyping an IoT-enabled alert system utilizing Raspberry Pi-based circuits. This system will allow the AI server to trigger visual or auditory signals when specific conditions are met, such as an overcrowded ferry or threshold-based event detection.
This server-to-device communication will enable automated responses in real-time, enhancing safety and operational efficiency. As development progresses, VASTIQ Solutions will be publishing a white paper detailing the methodologies and technological advancements behind VAST-NXT.ai, pending approval from the internal research team.
Comments
Post a Comment