That’s an ambitious project, but before we start coding we should think about safety and legal permissions for an abandoned metro. Once that’s cleared, I can design a lightweight app that logs coordinates, takes photos, and uses a depth sensor or LIDAR for wall contours. I’d probably lean on a cross‑platform framework like Flutter for UI, integrate with ARCore/ARKit for real‑time mapping, and store data in a local SQLite DB that syncs to the cloud if needed. Let me know what tech stack you’re comfortable with, and I’ll sketch out the architecture.

Alright, here’s a quick skeleton: 1. Front‑end: Flutter (or React Native if you prefer) • Screens: Home, Scan, Map View, Settings • Camera integration: use platform camera plugin for photo capture • AR: ARCore (Android) / ARKit (iOS) for depth data, wall contour extraction 2. Data layer • Local DB: SQLite (flutter sqflite) to store scan metadata, image paths, depth points • Sync: optional cloud sync via Firebase Firestore or custom REST API 3. Backend (optional) • Node.js/Express or Python Flask to receive uploads, perform heavy depth‑map processing if needed, and serve map tiles • Store processed maps in S3 or Cloud Storage 4. Architecture flow • User opens app → chooses “New Scan” → AR session starts • User walks around, app records depth frames, captures photos at intervals • Depth frames are processed on‑device to generate point cloud → simplified wall polygons • Data saved locally; on exit, user can export map, sync, or share 5. Extras • Offline mode: all data stays local, sync when network appears • Undo/redo for mis‑captured frames • Simple UI overlay with progress bar and compass • Log any anomalies (broken tiles, collapsed walls) with a quick “Report” button That’s the core. Let me know if you want deeper details on any layer or specific libraries.