Transportation planning and traffic operations studies increasingly rely on video data (CCTV, drone, roadside cameras, dashcams). However, extracting reliable mobility measures from video often requires repeated manual work and disconnected tools.
SimJam Computer Vision Analytics is an open-source application that turns raw traffic video into planning-ready outputs. It supports:
- Traffic object detection (cars, trucks, buses, bicycles, pedestrians, etc.)
- Multi-object tracking to keep consistent IDs over time
- Mobility analytics such as counts, trajectories, and speed estimation (when calibration is available)
- Exportable outputs (CSV summaries and structured results) to support planning studies and reporting
Typical use cases:
- Turning movement counts (TMC) and approach volumes
- Speed estimation and speed distributions
- Trajectory extraction for safety/near-miss analysis
- Before/after studies (traffic calming, signal timing, policy changes)
- Data preparation for microsimulation calibration/validation
A practical workflow is:
- Detect + track road users using YOLO-based models
Step 1 — Detection + Tracking
- Export analytics (counts / speeds / trajectories / summaries)
Step 2 — Analytics + Export
Short demo video - Click to Play
The easiest way to get started is to follow the step-by-step video tutorial:
🎥 Getting Started Tutorial (Click to Play)
Python version 3.12 and higher
Visual Studio code
This project is licensed under the MIT License. It uses Ultralytics YOLO which is licensed under AGPL-3.0. This project is distributed as open-source in compliance with that license.