Proactive Detection of Pothole and Walkable Path for Safe Mobility of Visually Challenged

Potholes are primary cause for road accidents; hence it is important to identify and classify the potholes using efficient Computer vision-based techniques. Potholes not only cause problems to car drivers but even pedestrians face problem while walking on a road having potholes. It is also a major problem for blind people who must be extremely careful while walking. This paper presents a machine learning and computer vision-based approach to identify potholes from an image and alert visually impaired person about the same. Dataset is compiled from various available open-source datasets along with few images clicked from smartphone. Five different classification algorithms are used for classification of the images after applying image pre-processing and feature vector extraction. The algorithm implemented in this paper uses machine learning based approach which is computationally efficient. It provides an accuracy at par with computationally expensive deep learning algorithms. The method involves feature extraction, dimensionality reduction and finally classification algorithms are applied. Binary classification algorithms are used to classify the images. Performance evaluation metrics like Precision, Recall, F1-Score and Accuracy are used to evaluate the model’s performance. Random forest model showed the best performance among all the algorithms with an accuracy of 88% on test dataset.