EVA-EAI-REST (Edge AI Restricted Area Detection) is a flagship-class, deep learning-based Edge Vision Analysis (EVA) software module designed to monitor unauthorized human or vehicle entry into prohibited areas in real time across facilities, critical infrastructure, industrial sites, and high-security environments.
Processing video streams with millisecond-level latency, EVA-EAI-REST continuously monitors digitally defined virtual security boundaries (ROI – Region of Interest) and restricted-area polygons, 24/7. As soon as an intrusion occurs, it distinguishes only object classes that represent an actual security threat (such as people, cars, trucks, and motorcycles) and generates a high-resolution event image (snapshot), video clip, and timestamp. It then delivers instant alerts to security and operations teams through multiple channels. Fully integrated with the EVC series Edge Vision Computer hardware family, this module is designed to provide the highest level of area security and perimeter protection.
Key Advantages and Capabilities
-
Object Classification-Based Area Intrusion Detection: Unlike conventional pixel-based motion detection systems, it recognizes human and vehicle shapes using deep learning models. By categorizing the target entering a restricted area (Human Intrusion / Vehicle Intrusion), it directly triggers the corresponding alarm scenario.
-
Multiple Virtual Polygons and Flexible Area Definition: Any number of restricted areas (ROI) with complex geometric shapes can be defined within the camera's field of view. Separate rules can be configured for each area, including object types, entry directions, and dwell times.
-
Intelligent False Alarm Filtering Against Environmental Factors: It reliably filters out environmental factors such as tree leaves swaying in the wind, heavy rain or snowfall, light reflections, cloud shadows, and stray animal movements, minimizing the false positive rate.
-
Time-Based (Scheduled) Security Rules: Dynamic schedules can be configured for areas that are accessible during working hours but restricted outside working hours or on weekends.
-
Zero Cloud Dependency and Continuous Edge AI Operation: Image analysis is performed locally and directly on the EVC (Edge Vision Computer) device at the site. It is unaffected by internet connectivity interruptions and provides zero bandwidth waste, minimal response times, and maximum data privacy.
-
Plug-and-Play Flexible Licensing: The module is provided as a ready-to-use component through the embedded EVC Runtime software architecture. Without requiring changes to the existing IP camera infrastructure, it can be activated and licensed with a single click for the required camera channels.
Architecture and Software Integration Capabilities
EVA-EAI-REST uses the GPU resources and hardware-based video decoders (Hardware H.264/H.265 Decode) of the EVC hardware family with optimal efficiency.
-
High Channel Density and Low Latency: It can simultaneously process high-resolution streams (including 4K/8MP) from dozens of camera channels on a single EVC device. It operates with high accuracy across thermal, infrared night-vision (IR), and standard optical cameras.
-
Extensive Integration Ecosystem (VMS and Perimeter Security Systems): Through REST API, WebSocket, MQTT, and dry-contact relay outputs, it integrates seamlessly with existing VMS (Video Management Software), barrier and turnstile systems, public address (PA) units, and centralized alarm panels.
Applications and Industry Use Cases
The EVA-EAI-REST module is an enterprise-grade solution designed for all indoor and outdoor environments where perimeter security, site monitoring, and unauthorized access prevention are critical:
-
Critical Infrastructure and Energy Facilities: Perimeters of power plants, transformer substations, solar and wind farms, dams, and water treatment facilities.
-
Industrial Facilities, Factories, and Heavy Industry: Hazardous machinery operating zones, robotic production cells, high-voltage rooms, and production lines restricted to authorized personnel.
-
Airports, Seaports, and Transportation Infrastructure: Runway perimeters, customs-controlled areas, railway lines, metro tunnels, and apron security zones.
-
Data Centers and Financial Institutions: Server rooms, vaults, and high-security control centers.
-
Military Zones, Public Buildings, and Logistics Warehouses: Border lines, warehouse loading ramps, and restricted facility areas outside working hours.
Comprehensive Technical Specifications
| Component / Module Feature | Technical Details |
|---|---|
| Module Name | EVA-EAI-REST (Edge AI Restricted Area Detection Module) |
| Primary Function | Real-time detection of unauthorized human and vehicle entry into prohibited areas |
| Detection Technology | Deep Learning, Real-Time Object Classification, and Position Tracking |
| Detectable Targets | People, Cars, Trucks / Buses, Motorcycles / Bicycles |
| Application Area | Area Security, Perimeter Protection, and Occupational Safety |
| Input Data | Live Camera Video Stream (RTSP, ONVIF Profile S/G/T, MJPEG, H.264, H.265) |
| Outputs and Notifications | Instant Pop-Up Image (Snapshot), Alarm Video Clip, Timestamped Alarm Log, Audible Siren / Relay Trigger |
| Response / Processing Time | < 300 ms (real-time area intrusion detection in under half a second) |
| Hardware Compatibility | All Edge Vision Computer hardware models in the EVC-04, EVC-08, EVC-16, EVC-24, EVC-32, EVC-40, EVC-48, EVC-56, and EVC-64 series |
| Operating Architecture | Edge AI (100% local and independent image processing directly on the device) |
| Software Integration | Fully integrated with the embedded EVC Runtime software and the EVA (Edge Vision Analysis) ecosystem |
| Area and Time Rules | Multiple Polygon ROI Definitions, Dwell Time, and Schedule-Based Activation |
| Integration Protocols | REST API, Webhook, MQTT, WebSocket, SDK, and Dry Contact (Relay) Output Support |
| Operating System Support | Windows 11 Pro for Workstations / Ubuntu LTS (on EVC Runtime) |
| Shop Location |
English
Türkçe
Русский