MahaRoadNet

The images were collected using smartphones (Samsung S23, Pixel 7, and Redmi Note 12) under diverse lighting conditions : sunny, overcast, shadows, and night with street lighting to capture the full variability of real-world road inspection scenarios. The data enables training and validation of object detection models such as YOLO and Faster R-CNN, as well as benchmarking of monocular depth estimation and pothole severity scoring algorithms.

The inclusion of varied lighting conditions and device types makes this dataset particularly valuable for developing robust mobile-based road inspection systems that generalize across smartphones and environmental contexts. Researchers and practitioners in computer vision, transportation engineering, and smart city applications will find this dataset useful for advancing automated road infrastructure monitoring.