The presence of potholes on the roads is one of the major causes of road accidents as well as wear and tear of vehicles. Various methods have been implemented to solve this problem ranging from manual reporting to authorities to the use of vibration-based sensors to 3D reconstruction using laser imaging. However, these methods have some limitations such as the high setup cost, risk while detection or no provision for night vision. In this work, we use the Mask R-CNN model to detect potholes, as it provides exceptional segmentation results. We synthetically generate a dataset for potholes, annotate it, do data augmentation and perform transfer learning on top of Mask R-CNN model which is pre-trained on MS COCO dataset. This support system was tested in varying lighting and weather conditions and was performed well in these situations as well.
CITATION STYLE
Kavati, I. (2022). Deep Learning-Based Pothole Detection for Intelligent Transportation Systems. In Lecture Notes in Electrical Engineering (Vol. 869, pp. 257–267). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-19-0019-8_20
Mendeley helps you to discover research relevant for your work.