🦓 ZebraRehab

ZebraRehab is a Python application that uses data from OpenStreetMap, orthophotos from the National Aerial Orthophotography Plan [PNOA], and a convolutional neural network (CNN) to identify, classify, and visualize pedestrian crosswalks in poor condition in the Valencian Community.

---

---

📑 Table of Contents

• Features
• Project structure
• Screenshots
• Installation
• Execution
• Technologies used
• Notes

---

🚀 Features

• 🔎 Data query (pedestrian crosswalks) from the Overpass API (OpenStreetMap)
• 🧩 Creation of regions of interest (ROIs) from the retrieved data
• 🚶 Extraction of aerial imagery of pedestrian crosswalks (PNOA)
• 🧠 Classification of crosswalk condition using a trained CNN model (test_acc: 84.9%)
• 🗺️ Visualization of results within the search region
• 💻 Interactive interface built with Streamlit

---

📁 Project structure

ZebraRehab/
│
├── data/
│   ├── icon/                 
│   ├── model/
│   │   └── CNN.h5            # Trained classification model
│   └── to_predict/           # Images pending classification
│
├── info/
│   ├── __init__.py
│   └── info.py               # Sidebar
│
├── Logica/
│   ├── __init__.py
│   ├── folders.py            # Folder and path management
│   ├── GetMap.py             # Image download/retrieval
│   ├── map.py                # Map rendering
│   ├── OSM.py                # Overpass API / OpenStreetMap queries
│   ├── predict.py            # Inference with the CNN model
│   └── ROIs.py               # Extraction of regions of interest (crosswalks)
│
├── misc/                     # Screenshots
├── main.py
├── requirements.txt
└── README.md

---

🖼️ Screenshots

📖 Usage instructions

📍 Regions of interest (ROIs)

🚶 Detected pedestrian crosswalks

📊 Classification results

---

⚙️ Installation

git clone https://github.com/jrvalza/ZebraRehab.git
cd ZebraRehab
pip install -r requirements.txt

💡 It is recommended to use a virtual environment (venv or conda) to avoid dependency conflicts.

python -m venv venv
source venv/bin/activate    # On Windows: venv\Scripts\activate
pip install -r requirements.txt

---

▶️ Execution

streamlit run main.py

The application will open automatically in your browser at http://localhost:8501.

---

🧠 Technologies used

Technology	Use
🐍 Python	Main language
🎈 Streamlit	User interface
🌐 Requests	HTTP requests to the Overpass API
🗺️ PNOA / OpenStreetMap / Overpass API	Geographic data source
🧠 TensorFlow / Keras	CNN model for crosswalk condition classification (CNN.h5)

---

📌 Notes

• The Overpass API may limit the number of requests if the service is overused.
• Analyzing smaller regions provides better performance when generating results.