🪸 Drone Imaging for Red Tide Detection (DIRTX)

DIRTX is a Flutter application that assists early detection of harmful algal blooms (red tide) by analyzing drone imagery and pre-recorded video using machine learning and computer-vision workflows. The app is a decision-support tool for marine biologists, environmental agencies, and local governments to visualize and review possible red tide clusters.

🔑 Key features

• Preview and analyze drone-captured images and videos.
• Visual overlay of detected HAB (harmful algal bloom) regions via instance segmentation (YOLOv9-based model in separate backend repo).
• Platform support: Linux, macOS, Windows, Android (via Flutter).

⚠️ Note

This repository contains the Flutter frontend only. The DIRTX Flask/YOLO backend lives in a separate repository: https://github.com/benny-18/DIRTX-YOLOv9-Model

🔗 Quick links

• Project: DIRTX
• Flutter entrypoint: lib/main.dart
• Backend connector: lib/misc/DIRTXFlaskBackend.dart
• LICENSE: LICENSE

🌟 Why use DIRTX

• Speeds up HABs detection using automated image analysis rather than manual water sampling.
• Portable UI for reviewing drone footage and marking suspected bloom areas.
• Designed as a complement to domain workflows — not a replacement for laboratory testing.

Getting started

⚙️ Prerequisites

• Install Flutter (stable channel). See https://flutter.dev/docs/get-started/install
• Ensure platform toolchains are set up for your target (Linux, macOS, Windows, Android).

📥 Clone and install

git clone git@github.com:benny-18/DIRTX.git
cd DIRTX
flutter pub get

▶️ Run the app (development)

• Run on desktop (Linux example):

flutter run -d linux

• Run on Android emulator or device:

flutter run -d <device-id>

📦 Build distributables

# Linux
flutter build linux

# Windows
flutter build windows

# macOS
flutter build macos

# Android APK
flutter build apk --release

💡 Usage examples

• Open the app and choose Input footage to load drone images or video.
• If connected to a running backend, trigger analysis from the footage viewer to send frames for model inference.
• Review detection overlays right after it is processed.

For UI-specific code paths, see the following key files:

• lib/DIRTXAppHome.dart — main app shell and navigation
• lib/DIRTXFootageSelectionCard.dart — media selection
• lib/misc/DIRTXOutputViewer.dart — results preview

❓ Getting help

• Open an issue in this repository for bugs or feature requests.
• For backend model issues or model training, see the backend repo: https://github.com/benny-18/DIRTX-YOLOv9-Model

👥 Maintainers

• Primary contact: Marvin Masubay (benny-18) (https://github.com/benny-18)

📄 License

This project is licensed under the terms in the LICENSE file.