Road Extraction Based on Improved Convolutional Neural Networks with Satellite Images
Deep learning has been applied in various fields for its effective and accurate feature learning capabilities in recent years. Currently, information extracted from remote sensing images with the learning methods has become the most relevant research area for its developed precision. In terms of dev...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-10-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/12/21/10800 |
| _version_ | 1797469276966223872 |
|---|---|
| author | Lei He Bo Peng Dan Tang Yuxia Li |
| author_facet | Lei He Bo Peng Dan Tang Yuxia Li |
| author_sort | Lei He |
| collection | DOAJ |
| description | Deep learning has been applied in various fields for its effective and accurate feature learning capabilities in recent years. Currently, information extracted from remote sensing images with the learning methods has become the most relevant research area for its developed precision. In terms of developing segmentation precision and reducing calculation power consumption, the improved deep learning methods have received more attention, and the improvement of semantic segmentation architectures has been a popular solution. This research presents a learning method named D-DenseNet with a new structure for road extraction. The methods for the improvement are divided into two stages: (1) alternate the consecutive dilated convolutions number in the structure of the network (2) the stem block is arranged as the initial block. So, dilated convolution can obtain more global context information through the whole network. Further, the D-DenseNet restructures D-LinkNet by taking DenseNet as its backbone instead of ResNet, which can expand the receptive field and accept more feature information. The D-DenseNet is effective because of its 119 M model size and 57.96% IoU on the processing test data and 99.3 M modes size and 66.26% on the public dataset, which achieved the research objective for reducing model size and developing segmentation precision—IoU. The experiment indicates that the D-Dense block and the stem block are effective for developing road extraction, and the appropriate number of convolution layers is also essential for model evaluation. |
| first_indexed | 2024-03-09T19:18:08Z |
| format | Article |
| id | doaj.art-6d687b322f49432bac41ad4141b72b2c |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-09T19:18:08Z |
| publishDate | 2022-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-6d687b322f49432bac41ad4141b72b2c2023-11-24T03:33:00ZengMDPI AGApplied Sciences2076-34172022-10-0112211080010.3390/app122110800Road Extraction Based on Improved Convolutional Neural Networks with Satellite ImagesLei He0Bo Peng1Dan Tang2Yuxia Li3School of Software Engineering, Chengdu University of Information Technology, Chengdu 610225, ChinaSchool of Automation, University of Electronic Science and Technology of China, Chengdu 610054, ChinaSchool of Software Engineering, Chengdu University of Information Technology, Chengdu 610225, ChinaSchool of Automation, University of Electronic Science and Technology of China, Chengdu 610054, ChinaDeep learning has been applied in various fields for its effective and accurate feature learning capabilities in recent years. Currently, information extracted from remote sensing images with the learning methods has become the most relevant research area for its developed precision. In terms of developing segmentation precision and reducing calculation power consumption, the improved deep learning methods have received more attention, and the improvement of semantic segmentation architectures has been a popular solution. This research presents a learning method named D-DenseNet with a new structure for road extraction. The methods for the improvement are divided into two stages: (1) alternate the consecutive dilated convolutions number in the structure of the network (2) the stem block is arranged as the initial block. So, dilated convolution can obtain more global context information through the whole network. Further, the D-DenseNet restructures D-LinkNet by taking DenseNet as its backbone instead of ResNet, which can expand the receptive field and accept more feature information. The D-DenseNet is effective because of its 119 M model size and 57.96% IoU on the processing test data and 99.3 M modes size and 66.26% on the public dataset, which achieved the research objective for reducing model size and developing segmentation precision—IoU. The experiment indicates that the D-Dense block and the stem block are effective for developing road extraction, and the appropriate number of convolution layers is also essential for model evaluation.https://www.mdpi.com/2076-3417/12/21/10800semantic segmentationlow-level roadinformation extractiondeep learning |
| spellingShingle | Lei He Bo Peng Dan Tang Yuxia Li Road Extraction Based on Improved Convolutional Neural Networks with Satellite Images Applied Sciences semantic segmentation low-level road information extraction deep learning |
| title | Road Extraction Based on Improved Convolutional Neural Networks with Satellite Images |
| title_full | Road Extraction Based on Improved Convolutional Neural Networks with Satellite Images |
| title_fullStr | Road Extraction Based on Improved Convolutional Neural Networks with Satellite Images |
| title_full_unstemmed | Road Extraction Based on Improved Convolutional Neural Networks with Satellite Images |
| title_short | Road Extraction Based on Improved Convolutional Neural Networks with Satellite Images |
| title_sort | road extraction based on improved convolutional neural networks with satellite images |
| topic | semantic segmentation low-level road information extraction deep learning |
| url | https://www.mdpi.com/2076-3417/12/21/10800 |
| work_keys_str_mv | AT leihe roadextractionbasedonimprovedconvolutionalneuralnetworkswithsatelliteimages AT bopeng roadextractionbasedonimprovedconvolutionalneuralnetworkswithsatelliteimages AT dantang roadextractionbasedonimprovedconvolutionalneuralnetworkswithsatelliteimages AT yuxiali roadextractionbasedonimprovedconvolutionalneuralnetworkswithsatelliteimages |