YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Pothole Detection using Deep Learning
This repository contains trained object detection models for pothole detection as part of a bachelor’s thesis project.
The repository is hosted on Hugging Face, with the corresponding source code and experiments available on GitHub:
👉 GitHub repository: https://github.com/DanielsStulpe/pothole-detection
📌 Overview
The project evaluates multiple deep learning architectures for automatic pothole detection in road surface images.
Models evaluated:
- YOLOv8
- YOLO11
- YOLO26
- Faster R-CNN
- RetinaNet
A final optimized YOLO26 model is also included.
🏆 Best Model
- Model: YOLO26 (optimized)
- File:
yolo26_best.pt
This model achieved the best overall performance across all experiments and hyperparameter optimizations.
📁 Model List
🔹 Baseline models
baseline_yolov8.ptbaseline_yolo11.ptbaseline_yolo26.ptbaseline_fasterrcnn.ptbaseline_retinanet.pt
🏆 Final model
yolo26_best.pt
📊 Task
- Object Detection
- Single-class (pothole detection)
📦 Usage (Ultralytics YOLO)
from ultralytics import YOLO
model = YOLO("yolo26_best.pt")
results = model("image.jpg")
results.show()
📚 Dataset
Dataset used in this project is publicly available on Roboflow:
https://universe.roboflow.com/bak-6b12k/pothole-idtb2
📄 License
This project is licensed under the MIT License.
MIT License
Copyright (c) 2026 Daniels Stulpe
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND.
🧠 Notes
This model is intended for research and academic purposes. Performance may vary depending on environmental conditions, dataset distribution, and image quality.