Project overview: Software for Infrastructure Objects Detection Using Machine Learning

This is my Bachelor's thesis at Petro Mohyla Black Sea National University, which consists of two parts:

• Training Building Segmentation Model
• Software Development

The software is a desktop application that detects infrastructure objects and displays them in a GUI built with PyQt6. It allows users to load satellite images, visualize detected features, and manage layer interactions.

The model is a U-Net with ResNet34 backbone which automatically detects building footprints from satellite images. The model performs semantic segmentation, identifying the exact pixels where buildings are located.

Repository Structure

model-train/ (use this in a separate venv)
├── building_segmentation.ipynb
├── requirements.txt
src/
├── gui/
├── object_detection/
├── resources/
│   ├── demo_images/
│   ├── icons/
│   └── model/
├── utils/
└── main.py
tests/
└── test_gui.py
app.spec
build.sh
requirements.txt
setup.py

Features

• Load and display satellite images
• Detect and overlay infrastructure objects
• Manage visual layers (add, remove, reorder)
• Simple and clean PyQt6 interface
• Modular design with separate logic and UI layers
• Includes automated tests for core GUI functionality
• Provides custom building segmentation model

How to Build and Run the App

Option 1: Run from Source

Clone the repository:

git clone https://github.com/MsRapunzel/software-for-infrastructure-objects-detection.git
cd software-for-infrastructure-objects-detection

Create and activate a virtual environment:

python -m venv venv
source venv/bin/activate

Install dependencies:

pip install -r requirements.txt

Run tests:

pip install pytest pytest-qt
PYTHONPATH=src pytest -s -v tests/test_gui.py

Run the application:

python src/main.py

Option 2: Build a Standalone Executable

Clone the repository:

git clone https://github.com/MsRapunzel/software-for-infrastructure-objects-detection.git
cd software-for-infrastructure-objects-detection

Make build.sh executable:

chmod +x build.sh

Run build script:

./build.sh

The built app will appear in the dist/ directory.

Option 3: Use a GitHub Release

1. Go to the Releases tab on GitHub (or open the latest).
2. Download .dmg for macOS.
3. Open the downloaded file, then drag and drop .app file to /Applications folder.

Note: On first launch macOS may block the app with warning: "macOS cannot verify that this app is free from malware". In this case:

1. Press and hold Control ^ and click the app, then select "Open" from the context menu.
2. Go to System Settings > Privacy & Security and click "Open Anyway" if the app has appeared under Security.