AI Occupancy Detection and Surveillance System


Transform any queue into a fast, frictionless experience with our AI-powered occupancy detection system. Using edge computing and real-time video analysis, it instantly identifies occupied positions and automatically updates LED screens with clear red (Occupied) and green (Available) indicators, helping customers find available spots quickly and dramatically reducing wait times.


Supports up to 39 camera channels and controls up to 11 LED displays, making it ideal for gas stations, drive-thru, retail checkouts, banks, and any high-traffic queue environment.

 

AI Driven Occupancy Detection System

 

Beyond gas station applications, we provide customized solutions for a wide range of industries and scenarios, including transportation, roadwork signage, traffic guidance, public facilities, and other specialized use cases. With flexible architecture and strong customization capability, we deliver tailored solutions to meet the needs of different industries and environments.

Advanced Capabilities with Intelligent Safety Monitoring


  • After-Hours Protection— Detects unauthorized people and vehicles during non-business hours while intelligently ignoring authorized staff in yellow vests.
  • Foreign Object Alerts— Spots unexpected items left in the station after closing.
  • Fire & Smoke Detection— Instantly identifies fire, smoke, or cigarette use for enhanced safety.
  • License Plate Recognition— Automatically reads and logs vehicle plates for better security and management.

AI-NVR for Safety and Surveillance Solutions


1.AI BOX

Option A: Nvidia Solutions

1 × Nvidia Jetson Orin AGX 32GB + 1 × Nvidia Jetson Orin NX 16GB


The system is built with one Nvidia Jetson Orin AGX 32GB and one Nvidia Jetson Orin NX 16GB. Based on the video decoding capability of the Jetson Orin AGX 32GB, the platform supports:

  • Up to 18 channels of 1080p @ 30 fps (H.265) with upgraded 2TB NVMe SSD

Jetson Orin AGX


Jetson Orin NX 16GB


Option B: AI inference and NVR processing

1 × One AI-PC (Intel Core Ultra 9 285H) + 1 × Dedicated GPU


AI-PC powered by the Intel Core Ultra 9 285H processor together with a dedicated GPU,  99 TOPS performance, 64GB memory, 1TB SSD, dual 10GbE ports, and expandable via EX Pro dock with NVIDIA RTX 50 Series GPU for massive computing power and storage.


AI-PC+ GPU


Key Benefits of Option B:

  • All-in-One Efficiency:Combines AI inference and NVR processing in a single powerful unit for a simpler system architecture.
  • Superior Performance:Delivers stronger video decoding with support for 24+ channels of 2K+ @ 30fps (H.265), ensuring higher stability.
  • Better Value & Flexibility:Offers excellent cost-to-performance ratio, richer I/O interfaces, and easier scalability.


2. IP Camera


  • SONY IMX335LQN-C STARVIS: Designed for high-definition daytime image quality and strong low-light performance at night.
  • Sensor: 6MP 1/2.8” low-light CMOS sensor; color: 0.0001 Lux @ F1.2; black-and-white: 0.0001 Lux @ F1.2
  • Video Processing: Advanced H.265+ encoding (compatible with H.265/H.264), dual-stream output, AVI format; adjustable bitrate from 0.1 Mbps to 8 Mbps; adjustable frame rate from 1 to 30 fps
  • Video Output:Main stream: 6MP 3072 × 2048 @ 20 fps, 5MP 2880 × 1616 @ 25 fps, 4MP 2560 × 1440 @ 25 fps, 2MP 1920 × 1080 @ 25 fps; Sub-stream: 800 × 448 @ 25 fps
  • Network Interface: One RJ45 Ethernet port, 10/100 Mbps adaptive; supports RTSP, FTP, DHCP, NTP, UPnP, and other network protocols
  • Protocol Support: ONVIF compliant for integration with third-party back-end systems and platforms
  • Power Supply: Supports PoE


3.LED Display

  • Pixel Pitch: 4.81 mm, 5 mm, and 4 mm
  • Dimensions:500 × 750 × 70 mm (P4.81)320 × 480 × 70 mm (P5)512 × 768 × 70 mm (P4)
  • Outdoor use: IP65 waterproof rating, US-standard plug, UL-certified power supply, power cable and network cable

LED

 (Any dimensions and functions can be customized to meet our client’s requirements.)


4.System Architecture and Workflow


System Architecture and Workflow:

Four Core System Modules


The system consists of four core modules: the front-end sensing layer, the edge intelligence layer, the network transmission layer, and the display interaction layer.


1.Front-End Sensing Layer

Multiple IP cameras are deployed to cover all fueling positions. These cameras periodically capture frames and transmit them to the AI box for analysis.


2.Edge Intelligence Layer

The edge intelligence layer is powered by the Jetson Orin NX 16GB, which serves as the core AI inference unit. By running an object detection model, the system analyzes video streams in real time and determines whether a vehicle is occupying a fueling position.


3.Network Transmission Layer

The network transmission layer uses multiple industrial-grade switches, including the TL-SG2216 and TL-IS200P-14G2F, to provide gigabit-speed data transmission between the IP cameras and the AI box. PoE is also supported to simplify system cabling and deployment. In addition, the USR-G806 industrial router provides dual-link backup through both 4G and wired connections, ensuring reliable remote maintenance and stable data upload.


4.Display Interaction Layer

The display interaction layer includes LED screens in various specifications, such as P5 320×480, P4.81 500×750, and P4 512×768. These screens work in coordination with the AI box to display system status in real time: red indicates “Occupied,” while green indicates “Available,” providing clear and intuitive guidance for customers.