What it does This is a curated markdown list of datasets for robotics and computer vision research. It organizes links by scenario (driving, flying, underwater, indoor, outdoor) and by technical topic (SLAM, tracking, 3D reconstruction, path planning). Think of it as a human-maintained index to the scattered, often buried data collections that labs publish and forget.

The interesting bit The curation is opinionated in a useful way: entries carry emoji signals (:+1: for recommended, :memo: for notes, strikethrough for dead links). That small editorial layer saves you from clicking through to a 404 on Radish or discovering too late that a “classic” dataset has rotted.

Key highlights

• Covers niche scenarios often omitted elsewhere: underwater (Marine Robotics Datasets), drone racing (UZH-FPV), planetary navigation (ASRL)
• Includes topic-specific corners like range-only localization and task-and-motion-planning benchmarks
• Tags research-group pages so you can follow prolific publishers (TUM CVG, Oxford VGG, Bonn Stachniss Lab)
• Explicitly marks broken links rather than pretending everything still works
• 511 stars suggests it’s a known bookmark, not a hidden gem

Caveats

• No search, no filtering, no metadata beyond the markdown: you scroll or you grep
• Some sections are thin (underwater gets one entry; path planning gets two)
• Last substantial update timing is unclear from the README

Verdict Grab this if you’re starting a new robotics project and need to survey what data exists before collecting your own. Skip it if you want a queryable database or automated download scripts — this is a reference page, not a tool.