Pavement Defect Segmentation in Orthoframes with a Pipeline of Three Convolutional Neural Networks
In the manuscript, the issue of detecting and segmenting out pavement defects on highway roads is addressed. Specifically, computer vision (CV) methods are developed and applied to the problem based on deep learning of convolutional neural networks (ConvNets). A novel neural network structure is con...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-08-01
|
| Series: | Algorithms |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-4893/13/8/198 |
| _version_ | 1827709676467257344 |
|---|---|
| author | Roland Lõuk Andri Riid René Pihlak Aleksei Tepljakov |
| author_facet | Roland Lõuk Andri Riid René Pihlak Aleksei Tepljakov |
| author_sort | Roland Lõuk |
| collection | DOAJ |
| description | In the manuscript, the issue of detecting and segmenting out pavement defects on highway roads is addressed. Specifically, computer vision (CV) methods are developed and applied to the problem based on deep learning of convolutional neural networks (ConvNets). A novel neural network structure is considered, based on a pipeline of three ConvNets and endowed with the capacity for context awareness, which improves grid-based search for defects on orthoframes by considering the surrounding image content—an approach, which essentially draws inspiration from how humans tend to solve the task of image segmentation. Also, methods for assessing the quality of segmentation are discussed. The contribution also describes the complete procedure of working with pavement defects in an industrial setting, involving the workcycle of defect annotation, ConvNet training and validation. The results of ConvNet evaluation provided in the paper hint at a successful implementation of the proposed technique. |
| first_indexed | 2024-03-10T17:25:15Z |
| format | Article |
| id | doaj.art-56a07e57368f4904bd6b233072aa1351 |
| institution | Directory Open Access Journal |
| issn | 1999-4893 |
| language | English |
| last_indexed | 2024-03-10T17:25:15Z |
| publishDate | 2020-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Algorithms |
| spelling | doaj.art-56a07e57368f4904bd6b233072aa13512023-11-20T10:11:54ZengMDPI AGAlgorithms1999-48932020-08-0113819810.3390/a13080198Pavement Defect Segmentation in Orthoframes with a Pipeline of Three Convolutional Neural NetworksRoland Lõuk0Andri Riid1René Pihlak2Aleksei Tepljakov3Department of Software Science, Tallinn University of Technology, 19086 Tallinn, EstoniaDepartment of Software Science, Tallinn University of Technology, 19086 Tallinn, EstoniaDepartment of Software Science, Tallinn University of Technology, 19086 Tallinn, EstoniaDepartment of Computer Systems, Tallinn University of Technology, 19086 Tallinn, EstoniaIn the manuscript, the issue of detecting and segmenting out pavement defects on highway roads is addressed. Specifically, computer vision (CV) methods are developed and applied to the problem based on deep learning of convolutional neural networks (ConvNets). A novel neural network structure is considered, based on a pipeline of three ConvNets and endowed with the capacity for context awareness, which improves grid-based search for defects on orthoframes by considering the surrounding image content—an approach, which essentially draws inspiration from how humans tend to solve the task of image segmentation. Also, methods for assessing the quality of segmentation are discussed. The contribution also describes the complete procedure of working with pavement defects in an industrial setting, involving the workcycle of defect annotation, ConvNet training and validation. The results of ConvNet evaluation provided in the paper hint at a successful implementation of the proposed technique.https://www.mdpi.com/1999-4893/13/8/198pavement distressdigital imageDeep Learningtransfer learningConvolutional Neural Networkactive learning |
| spellingShingle | Roland Lõuk Andri Riid René Pihlak Aleksei Tepljakov Pavement Defect Segmentation in Orthoframes with a Pipeline of Three Convolutional Neural Networks Algorithms pavement distress digital image Deep Learning transfer learning Convolutional Neural Network active learning |
| title | Pavement Defect Segmentation in Orthoframes with a Pipeline of Three Convolutional Neural Networks |
| title_full | Pavement Defect Segmentation in Orthoframes with a Pipeline of Three Convolutional Neural Networks |
| title_fullStr | Pavement Defect Segmentation in Orthoframes with a Pipeline of Three Convolutional Neural Networks |
| title_full_unstemmed | Pavement Defect Segmentation in Orthoframes with a Pipeline of Three Convolutional Neural Networks |
| title_short | Pavement Defect Segmentation in Orthoframes with a Pipeline of Three Convolutional Neural Networks |
| title_sort | pavement defect segmentation in orthoframes with a pipeline of three convolutional neural networks |
| topic | pavement distress digital image Deep Learning transfer learning Convolutional Neural Network active learning |
| url | https://www.mdpi.com/1999-4893/13/8/198 |
| work_keys_str_mv | AT rolandlouk pavementdefectsegmentationinorthoframeswithapipelineofthreeconvolutionalneuralnetworks AT andririid pavementdefectsegmentationinorthoframeswithapipelineofthreeconvolutionalneuralnetworks AT renepihlak pavementdefectsegmentationinorthoframeswithapipelineofthreeconvolutionalneuralnetworks AT alekseitepljakov pavementdefectsegmentationinorthoframeswithapipelineofthreeconvolutionalneuralnetworks |