Overview
Intel OpenVINO™ Smart Intersection Analytics
Short summary: Smart Intersection is an AI-powered edge video analytics solution for intelligent transportation and smart city scenarios. It fuses multi-camera feeds to track vehicles and pedestrians, estimate speed and direction, and provide intersection-level analytics optimized with Intel OpenVINO pipelines.
About Advantech Container Catalog (ACC)
Advantech Container Catalog is a comprehensive collection of ready-to-use, containerized software packages designed to accelerate the development and deployment of Edge AI applications. By offering pre-integrated solutions optimized for embedded hardware, it simplifies the challenges often faced with software and hardware compatibility, especially in GPU/NPU-accelerated environments.
| Feature / Benefit | Description |
|---|---|
| Accelerated Edge AI Development | Ready-to-use containerized solutions for faster prototyping and deployment |
| Hardware Compatible | Reduces hardware and package incompatibility issues |
| GPU/NPU Access Ready | Supports passthrough for efficient hardware acceleration |
| Model Conversion & Optimization | Built-in model conversion and quantization recommendations |
| Optimized for CV & LLM Applications | Optimized stacks for vision and language workloads |
Container Overview
This project packages Intel's Smart Intersection reference application into a containerized deployment that orchestrates DL Streamer pipelines, SceneScape services, Grafana, InfluxDB, MQTT, and Node-RED. It enables multi-camera ingestion, OpenVINO-accelerated inference (CPU/GPU/NPU), fused tracking, scene analytics and dashboard visualization for intersection-level traffic monitoring.
Demo
Use Case
- Real-time multi-camera vehicle and pedestrian tracking at intersections
- Traffic behavior analysis: speed, direction, interactions, and ROI-based metrics
- Edge deployment for smart city traffic monitoring and incident detection
- Integration with dashboards (Grafana) and downstream analytics pipelines
Key Features
- End-to-end Smart Intersection reference stack (DL Streamer, SceneScape, Grafana, InfluxDB)
- OpenVINO-optimized inference with selectable device backends (CPU / iGPU / NPU)
- Prebuilt dashboards and SceneScape UI for visualizing fused tracks and analytics
Host Device Prerequisites
| Item | Specification |
|---|---|
| Compatible Hardware | x86 server or edge platform with CPU; optional Intel iGPU / NPU for acceleration |
| Platform Version | Linux host supported (refer to resources/system-requirements.md) |
| Host OS | Linux with Docker Engine and Docker Compose |
| Required Packages | Git, Docker Engine, Docker Compose, OpenVINO runtime (for GPU/NPU) |
| Software Installation Guide | resources/system-requirements.md and repository README.md |
Required Software Packages on Host Device
| Component | Version | Description |
|---|---|---|
| Docker Engine | Latest supported | Container runtime |
| Docker Compose | Latest supported | Service orchestration |
| OpenVINO | Matching release (e.g., 2026.x) | GPU/NPU inference runtime and pipelines |
| Intel compute runtime / NPU drivers | As required | Device-specific drivers for iGPU/NPU |
Container Environment Overview
Software Components in the Image
| Component | Version | Description |
|---|---|---|
| DL Streamer Pipeline Server | repo-provided | Video ingestion and inference pipeline host |
| Intel SceneScape Services | repo-provided | Scene management and fused tracking UI |
| Grafana | included | Dashboard visualization |
| InfluxDB | included | Time-series store for analytics |
| Node-RED | included | Integration and orchestration flows |
| MQTT Broker | included | Telemetry and messaging bus |
Container Quick Start Guide
For container quick start, including the docker-compose file and more, please refer to Advantech Container Github Repository
Supported AI Capabilities
Vision Models
| Model Family | Versions | Notes |
|---|---|---|
| Object Detection / Tracking | OpenVINO models | Used in DL Streamer pipelines for detection and tracking |
| Analytics / ROI metrics | Custom scenes | Fused tracking and ROI-based analytics in SceneScape |
Language Models Recommendation
| Model Family | Suggested Sizes | Memory Req. |
|---|---|---|
| N/A | N/A | N/A |
Optimization tips: Use the OpenVINO conversion and optimization guides in resources/ to convert and optimize models for GPU/NPU backends.
Supported AI Model Formats
| Format | Support Level | Notes |
|---|---|---|
| OpenVINO IR / OMZ | Full | Native runtime and optimized pipelines |
| ONNX | Partial | Convert to OpenVINO for best performance |
Hardware Acceleration Support
| Accelerator | Support Level | Compatible Libraries | Notes |
|---|---|---|---|
| CPU | Full | OpenVINO CPU plugins | Default inference device |
| Intel iGPU | Optional | OpenVINO GPU plugin, compute-runtime | See resources/how-to-use-gpu-for-inference.md |
| Intel NPU | Optional | NPU drivers and OpenVINO NPU plugin | See resources/how-to-use-npu-for-inference.md |
Troubleshooting & Notes
- Common issues:
- Missing device drivers for iGPU/NPU — verify kernel modules and vendor drivers.
- Docker post-installation: configure non-root Docker access if permission denied.
- Application-specific service failures — check container logs (
docker compose logs).
- Known limitations:
- Out-of-the-box configuration runs on CPU; GPU/NPU acceleration requires driver and OpenVINO runtime setup.
- Some resources and demos assume local assets installed via
install.sh.
References
- Intel Open Edge Platform: https://docs.openedgeplatform.intel.com/2026.0/OEP-articles.html
- Metro AI Suite: https://builders.intel.com/intel-technologies/software/edge-ai-suites/metro-ai-suite
- Smart Intersection reference: https://docs.openedgeplatform.intel.com/2026.0/edge-ai-suites/smart-intersection/index.html