Success case
Streamlining Transportation with Jetson: Advanced Multi-category and Cross-lens Tracking for Complex Scenarios
Hardware: NVIDIA Jetson Xavier NX Application: Object Tracking for Multiple Categories across Cameras Industry: Smart Transportation Deployment Location: US
- Industry
- Transportation & Logistics
- Application
- Cross-camera traffic tracking
Project background and needs
This cross-camera traffic tracking solution targets US smart-transportation scenarios, using Jetson Xavier NX with multi-category detection, re-identification, and trajectory fusion to keep vehicle and pedestrian tracking continuous.
Multi-category targets
- Traffic systems need to distinguish cars, trucks, bicycles, and similar types.
- Pedestrian identification and trajectory monitoring also matter for traffic safety.
Needs simultaneous multi-category target detection and tracking.
Cross-camera continuity
- Viewpoint changes and occlusion between cameras add tracking difficulty.
- Data from multiple cameras also needs to be coordinated and fused.
Needs re-identification combined with trajectory fusion to maintain tracking.
Solution and architecture
Detection and single-camera tracking
Jetson Xavier NX runs a module with an HRNetV2-W18 backbone, supporting single- and multi-category target tracking.
Cross-camera fusion
Built on DeepSORT, combining detection, ReID, and trajectory fusion to connect target tracks across camera switches.
Solution components
See the solution in action
On-site gallery (1)
Discuss your application with our team
Tell us about your requirements and deployment environment.