Logo Lanfrica
  • Home
  • Atlas
  • Insights
  • Docs
  • Sign in

© 2026 Lanfrica. All rights reserved. All copyrights of the resources shown on the Lanfrica website belong to the original copyright holders, unless explicitly stated otherwise.

hatieku-boateng/road-defect-collector

Domain:

mobility

Record type:

software
Creator:
hat
Host:
Mobile-friendly road image collection and review system for Ghana road defect monitoring. # Ghana Road Defect Monitoring A mobile-friendly road image collection and quality-review system designed to prepare reliable Ghanaian road-defect data for later AI modelling. ## Features - Collector ID and road-area capture - Manual camera capture or road image upload - Device GPS capture with accuracy and timestamp - Final record review before submission - Private image storage with Vercel Blob - Metadata storage with Neon Postgres - Password-protected administrator dashboard - Submission totals and filters - OpenStreetMap location overview - Approve or reject quality-review workflow - Local drone-video frame extraction - Open-source Grounding DINO pothole candidate detection - Human selection of AI candidates before upload ## Application routes - `/` — project landing page - `/collect` — collector submission form - `/drone` — browser-based drone AI analysis - `/admin/login` — administrator sign-in - `/admin` — protected review dashboard - `/api/health` — deployment configuration check ## Required environment variables Copy `.env.example` to `.env.local` and configure: ```text DATABASE_URL BLOB_READ_WRITE_TOKEN ADMIN_PASSWORD SESSION_SECRET ``` The database table and indexes are created automatically on the first database operation. ## Drone AI pilot The drone workflow samples up to 12 frames at two-second intervals. The Apache-2.0 Grounding DINO Tiny ONNX model runs in the reviewer's browser through Transformers.js. The original video is not uploaded. Only candidate frames chosen by a human reviewer are saved, and they enter the normal administrator review queue with a pending status. The model is downloaded from Hugging Face on first use and cached by the browser. This pilot searches only for potholes and must not be treated as a final trained road-defect model. ## Local development ```bash npm install npm run dev ``` Open `localhost` in a browser. ## Deploy a preview to Vercel ```bash vercel ``` Only promote to production after rev …

Visit

github.com

Tasks

computer vision

Similar

DiyeMark/Road-Defect-Detection-Systemlouicoder/Plastic-CollectorAalaa99k/oromo-corpus-collectorephantuk/kikuyu-language-collectorDatsede04/Amharic-corps-collector-botDatsede04/amharic-voice-data-collector-bot

DiyeMark/Road-Defect-Detection-System

A road defect detection system for Addis Ababa, Ethiopia # Road-Defect-Detection-System A Road Defe

louicoder/Plastic-Collector

A Mobile app for the project towards monitoring plastic collection in the country [ Uganda ] # Plas

Aalaa99k/oromo-corpus-collector

Production-ready clipboard corpus collector for Afaan Oromo. Automatically cleans, detects, deduplic

ephantuk/kikuyu-language-collector

Language collection for AI training

Datsede04/Amharic-corps-collector-bot

A Telegram Bot for Amharic Speech Data Collection # Lesan crowd sourcer bot @LesanEarBot A Telegr

Datsede04/amharic-voice-data-collector-bot

This project aims to collect and create an open source amharic data corpes for amharic speach recogn