Skip to content

About

Mangrove canopy detection system using computer vision and Detectron2

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

146 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MangroVision

Setting up on another laptop? Start with the step-by-step groupmate setup guide. It lists what to clone, what must be copied separately, and how to check that the app runs.

MangroVision is a FastAPI and React mangrove-planning system. It analyzes geotagged drone imagery, keeps eroded-area planting points visible but dynamically unavailable, manages planting assignments, and records aggregate organization monitoring visits for LGU staff.

MangroVision opens at /. The LIKE public website and planting appointment intake are available separately at /landing-page. See LIKE website development and integration for preview instructions, placeholders, and the separate development database setup.

Supported production application

Only these applications are supported as production writers:

  • MangroVision_New/api: FastAPI backend and the only database/storage client.
  • MangroVision_New/client: React frontend using credentialed HTTP-only cookie sessions.

Production persistence uses PostgreSQL/PostGIS and private S3-compatible object storage. SQLite and mutable zone GeoJSON files are migration inputs only. See Production database operations for local setup, migration, cutover, backup, restore, and Supabase deployment instructions.

For a groupmate's laptop, follow the clone and setup guide. The current orthophoto GeoTIFF and preferred canopy checkpoint are shared separately; Git does not contain the image data, private credentials, or those two large files.

Local development

  1. Copy .env.example to .env and replace every placeholder secret.

  2. Start Docker Desktop.

  3. Start PostgreSQL/PostGIS and MinIO:

    docker compose up -d
  4. Install Python dependencies and upgrade the database:

    python -m pip install -r requirements.txt
    alembic upgrade head
  5. Create the first administrator using explicit BOOTSTRAP_ADMIN_* values:

    python scripts/bootstrap_admin.py
  6. Install the frontend dependencies, then start the managed development server from the repository root:

    npm --prefix MangroVision_New/client ci
    python MangroVision_New/start_dev.py

    This starts FastAPI on port 8000 and React on port 5173. Running the launcher again restarts this workspace's existing managed stack. Ctrl+C stops both servers; on Windows, closing or terminating the launcher also removes its child processes. To stop the stack from another terminal:

    python MangroVision_New/start_dev.py --stop

    Use the launcher for both services. Independently launched servers and other programs occupying these ports are reported without being terminated.

Readiness is exposed at /api/health/ready and fails when PostgreSQL, Alembic, or private object storage is unavailable.

Important behavior

  • Images without GPS or fully outside active gis_coverage zones are rejected before analysis. Partial overlap requires explicit user confirmation.
  • Eroded zones never delete or hide planting points. Covered points display as Not Available for Planting; removing the zone dynamically returns eligible points to Planned.
  • Monitoring is recorded once per organization visit: alive/dead totals, average height, overall health, and LGU actions. No photo is required.
  • Analysis images are private objects and are returned only through short-lived signed URLs; image bytes are never stored in PostgreSQL.

About

Mangrove canopy detection system using computer vision and Detectron2

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages