Pothole Detection using Lidar

This project is a unified Flask-based web application that provides access to three advanced LiDAR road surface analysis methods: 1. Auto Correlation Method: Uses Linear Predictive Coding (LPC) for road roughness analysis, enhanced with YOLO-based visual confirmation. 2. Heatmap Method: Generates visual heatmaps of road surfaces using DBSCAN clustering, Savitzky-Golay filtering, and YOLO defect detection, with integrated folder management. 3. Random Forest Method: Employs machine learning (Random Forest) for automated road defect classification and analysis.