mahdichouchen/road-damage-tunisia-ai

# Road Damage Detection – Tunisia 🇹🇳 ## Overview This project focuses on detecting and classifying road surface damage in Tunisia using computer vision and deep learning. Poor road quality affects safety, vehicle maintenance costs, and transportation efficiency. The goal of this project is to support data-driven road maintenance decisions. --- ## Damage Types The system detects the following road damage categories: - 🕳 Potholes - ➖ Longitudinal cracks - ➖ Transverse cracks - 🕸 Alligator cracks --- ## Dataset This project uses the **RDD-2022 (Road Damage Detection)** dataset. Dataset structure: - Train / Validation / Test splits - Each split contains `images/` and `labels/` - Annotations are provided as bounding boxes --- ## Task Definition - **Task:** Object detection & classification - **Input:** Road images - **Output:** Bounding boxes + damage class labels --- ## Project Structure See the repository structure for data, configuration files, source code, and outputs. --- ## Status Project setup and configuration completed. Model training and evaluation in progress.