An Effective Model for Smartphone Based Pothole Classification and Admin Alerting System

This research analyses and compares machine learning and deep learning models for classifying pothole images and recording exact GPS coordinates using smartphones' built-in sensors. This deep learning classification algorithm uses convolutional neural networks (CNN) built with the TensorFlow library. Five distinct deep learning and machine learning models were trained and compared (SVM, Random Forest, Decision trees, KNN, InceptionResnetV2, InceptionV3). The most accurate model has a 91 percent accuracy rate. According to the results of this study, the suggested approach is reliable and inexpensive.