Abstract
The article is devoted to a comparative analysis of the effectiveness of convolutional neural networks for semantic segmentation of road surface damage marking. Currently, photo and video surveillance methods are used to control the condition of the road surface. Assessing and analyzing new manual data can take too long. Thus, a completely different procedure is required to inspect and assess the state of controlled objects using technical vision. The authors compared 3 neural networks (Unet, Linknet, PSPNet) used in semantic segmentation using the example of the Crack500 dataset. The proposed architectures have been implemented in the Keras and TensorFlow frameworks. The compared models of neural convolutional networks effectively solve the assigned tasks even with a limited amount of training data. High accuracy is observed. The considered models can be used in various segmentation tasks. The results obtained can be used in the process of modeling, monitoring, and predicting the wear of the road surface.
Highlights
Continuous video monitoring of the road surface can be an extremely tedious task for humans, but a straightforward task for automated computer vision (CV) systems
The results of predictions of semantic fracture extraction are shown for various Backbones in the case of the U-Net neural network in Table
U-Net neural network The results of predictions of semantic fracture isolation are shown for various Backbones in the case of the U-Net neural network in Table I.: Backbones vgg16 resnet18 seresnet18 resnext50 seresnext50 senet154 densenet121 inceptionv3 inceptionresnetv2 mobilenet2 efficientnetb0
Summary
Continuous video monitoring of the road surface can be an extremely tedious task for humans, but a straightforward task for automated computer vision (CV) systems. As noted in [1], transport infrastructure is the basis of the national economy, which needs to be systematically improved. Many researches are devoted to improving algorithms for detecting road defects. Depending on the method of road surface monitoring, defects can be detected both on two-dimensional images (2D) and on three-dimensional (3D) images [2] obtained by laser scanning in the form of a point cloud. Compared to two-dimensional (2D) pavement images, three-dimensional pavement data is less vulnerable to lighting conditions and provides more useful information. 2D methods cannot detect some defects due to the lack of depth information
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
