opendatacam/opendatacam

Health v2 · maintenance only

Adoption not part of the score

1725 stars · 300 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

OpenDataCam is an open-source computer vision application that detects and tracks moving objects in camera feeds or video files using YOLO/Darknet, counting them via user-defined virtual counters. It runs self-contained on NVIDIA GPUs and Jetson edge devices, exposing a web UI and REST API while keeping all data local.

Use cases

• count cars and pedestrians from a camera for traffic studies
• measure modal split and turn counts at intersections
• analyze pedestrian trajectories in a public square
• run object counting on a Jetson Nano at the edge
• analyze pre-recorded video files for moving object counts
• build a smart city sensor with privacy-preserving on-device analytics

When to choose

• you need to count or track moving objects (vehicles, bikes, pedestrians) from video without saving footage
• you want a self-contained, privacy-friendly edge deployment on NVIDIA Jetson or GPU servers
• you need an API-driven counting tool with a ready-made web UI and trajectory analysis

When to avoid

• you need object detection on CPU-only hardware or non-NVIDIA GPUs
• you need to store or analyze the video imagery itself, since OpenDataCam discards visual data
• you need a general-purpose video surveillance or recording system

Facets

application · maturity active

computer-vision video-processing analytics data-visualization api-framework computer-vision iot analytics machine-learning self-hosted yolo darknet jetson object-tracking edge-computing traffic-counting smart-city video-analytics privacy-preserving traffic-analysis docker linux gpu web-server

Member repositories

For agents

markdown · JSON · MCP: product_card(name="opendatacam/opendatacam")