Efficient road infrastructure is pivotal for ensuring the safety and smooth operation of urban environments. Potholes, often caused by natural or human-induced factors, pose significant threats to vehicular safety and traffic flow. Timely detection of potholes is crucial for proactive maintenance and prevention of accidents. In recent years, computer vision techniques have emerged as promising tools for automated pothole detection. This research focuses on enhancing existing vision-based methods using a combination of the Scale-Invariant Feature Transform (SIFT) and Principal Component Analysis (PCA) algorithms. SIFT, a cornerstone in feature extraction, is employed to identify distinctive features within images of road surfaces. To streamline subsequent processing and improve accuracy, the complementary PCA is utilized to reduce the dimensionality of the feature descriptors generated by SIFT. The resulting feature vectors are then fed into a Support Vector Machine (SVM) classifier for training and pothole classification. To evaluate the performance of the trained model, a Receiver Operating Characteristic (ROC) curve was plotted. The Area Under the Curve (AUC) was calculated to be 92%. Furthermore, the overall accuracy of the system was found to be 94.7%. Experimental results demonstrate that the combined use of SIFT and PCA outperforms the SIFT algorithm alone in terms of pothole detection accuracy. This finding underscores the efficacy of the proposed approach in addressing the challenges associated with automated pothole detection. The developed system offers a potential solution for cost-effective and timely road maintenance, contributing to safer and more efficient transportation systems.

Pothole Detection Scale-Invariant Feature Transform (SIFT) Support vector machine (SVM) Principal component analysis (PCA) vision-based methods