Detecting Floodwater on Roadways from Image Data Using Mask-R-CNN

Abstract

Object detection and segmentation algorithms evolved significantly in the last decade. Simultaneous object detection and segmentation paved the way for real-time applications such as autonomous driving. Detection and segmentation of (partially) flooded roadways are important inputs for vehicle routing and traffic management systems. This paper proposes an automatic floodwater detection and segmentation method utilizing the Mask-R-CNN algorithm - a deep learning algorithm belonging to Region-Based Convolutional Neural Networks (R-CNN) family of models for object detection and semantic segmentation. As the latest evolution in the R-CNN family, Mask-R-CNN fuses localization, classification, and segmentation in a compact and fast algorithm. To train the model, manually labeled images with urban, suburban, and natural settings are used. The performance of the algorithm is assessed in accurately detecting the floodwater captured in images. The results show that the proposed floodwater detection and segmentation perform better than previous studies.