A comparison of deep learning algorithms on image data for detecting floodwater on roadways

Object detection and segmentation algorithms evolved significantly in the last decade. Simultaneous object paved way for real-time applications such as autonomous driving. Detection of (partially) flooded roadways are essential inputs vehicle routing traffic management systems. This paper proposes an automatic floodwater method utilizing Mask Region-Based Convolutional Neural Networks (Mask-R-CNN) Generative Adversarial (GAN) algorithms. To train model, manually labeled images with urban, suburban, natural settings used. The performances assessed accurately detecting captured images. results show that proposed Mask-R-CNN-based outperform previous studies, whereas GAN-based model has a straightforward implementation compared to other models.