Building Outline Extraction Directly Using the U<sup>2</sup>-Net Semantic Segmentation Model from High-Resolution Aerial Images and a Comparison Study
Deep learning techniques have greatly improved the efficiency and accuracy of building extraction using remote sensing images. However, high-quality building outline extraction results that can be applied to the field of surveying and mapping remain a significant challenge. In practice, most buildin...
Main Authors: | , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-08-01
|
Series: | Remote Sensing |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-4292/13/16/3187 |
_version_ | 1797522172999106560 |
---|---|
author | Xinchun Wei Xing Li Wei Liu Lianpeng Zhang Dayu Cheng Hanyu Ji Wenzheng Zhang Kai Yuan |
author_facet | Xinchun Wei Xing Li Wei Liu Lianpeng Zhang Dayu Cheng Hanyu Ji Wenzheng Zhang Kai Yuan |
author_sort | Xinchun Wei |
collection | DOAJ |
description | Deep learning techniques have greatly improved the efficiency and accuracy of building extraction using remote sensing images. However, high-quality building outline extraction results that can be applied to the field of surveying and mapping remain a significant challenge. In practice, most building extraction tasks are manually executed. Therefore, an automated procedure of a building outline with a precise position is required. In this study, we directly used the U<sup>2</sup>-net semantic segmentation model to extract the building outline. The extraction results showed that the U<sup>2</sup>-net model can provide the building outline with better accuracy and a more precise position than other models based on comparisons with semantic segmentation models (Segnet, U-Net, and FCN) and edge detection models (RCF, HED, and DexiNed) applied for two datasets (Nanjing and Wuhan University (WHU)). We also modified the binary cross-entropy loss function in the U<sup>2</sup>-net model into a multiclass cross-entropy loss function to directly generate the binary map with the building outline and background. We achieved a further refined outline of the building, thus showing that with the modified U<sup>2</sup>-net model, it is not necessary to use non-maximum suppression as a post-processing step, as in the other edge detection models, to refine the edge map. Moreover, the modified model is less affected by the sample imbalance problem. Finally, we created an image-to-image program to further validate the modified U<sup>2</sup>-net semantic segmentation model for building outline extraction. |
first_indexed | 2024-03-10T08:25:42Z |
format | Article |
id | doaj.art-bfe94fa43cd04a10980e60e5a01229e5 |
institution | Directory Open Access Journal |
issn | 2072-4292 |
language | English |
last_indexed | 2024-03-10T08:25:42Z |
publishDate | 2021-08-01 |
publisher | MDPI AG |
record_format | Article |
series | Remote Sensing |
spelling | doaj.art-bfe94fa43cd04a10980e60e5a01229e52023-11-22T09:33:28ZengMDPI AGRemote Sensing2072-42922021-08-011316318710.3390/rs13163187Building Outline Extraction Directly Using the U<sup>2</sup>-Net Semantic Segmentation Model from High-Resolution Aerial Images and a Comparison StudyXinchun Wei0Xing Li1Wei Liu2Lianpeng Zhang3Dayu Cheng4Hanyu Ji5Wenzheng Zhang6Kai Yuan7School of Geography, Geomatics and Planning, Jiangsu Normal University, Xuzhou 221116, ChinaSchool of Geography, Geomatics and Planning, Jiangsu Normal University, Xuzhou 221116, ChinaSchool of Geography, Geomatics and Planning, Jiangsu Normal University, Xuzhou 221116, ChinaSchool of Geography, Geomatics and Planning, Jiangsu Normal University, Xuzhou 221116, ChinaSchool of Mining and Geomatics, Hebei University of Engineering, Handan 056038, ChinaSchool of Geography, Geomatics and Planning, Jiangsu Normal University, Xuzhou 221116, ChinaSchool of Geography, Geomatics and Planning, Jiangsu Normal University, Xuzhou 221116, ChinaSchool of Geography, Geomatics and Planning, Jiangsu Normal University, Xuzhou 221116, ChinaDeep learning techniques have greatly improved the efficiency and accuracy of building extraction using remote sensing images. However, high-quality building outline extraction results that can be applied to the field of surveying and mapping remain a significant challenge. In practice, most building extraction tasks are manually executed. Therefore, an automated procedure of a building outline with a precise position is required. In this study, we directly used the U<sup>2</sup>-net semantic segmentation model to extract the building outline. The extraction results showed that the U<sup>2</sup>-net model can provide the building outline with better accuracy and a more precise position than other models based on comparisons with semantic segmentation models (Segnet, U-Net, and FCN) and edge detection models (RCF, HED, and DexiNed) applied for two datasets (Nanjing and Wuhan University (WHU)). We also modified the binary cross-entropy loss function in the U<sup>2</sup>-net model into a multiclass cross-entropy loss function to directly generate the binary map with the building outline and background. We achieved a further refined outline of the building, thus showing that with the modified U<sup>2</sup>-net model, it is not necessary to use non-maximum suppression as a post-processing step, as in the other edge detection models, to refine the edge map. Moreover, the modified model is less affected by the sample imbalance problem. Finally, we created an image-to-image program to further validate the modified U<sup>2</sup>-net semantic segmentation model for building outline extraction.https://www.mdpi.com/2072-4292/13/16/3187building edge extractionhigh resolution imagesemantic segmentationedge detectiondeep learning |
spellingShingle | Xinchun Wei Xing Li Wei Liu Lianpeng Zhang Dayu Cheng Hanyu Ji Wenzheng Zhang Kai Yuan Building Outline Extraction Directly Using the U<sup>2</sup>-Net Semantic Segmentation Model from High-Resolution Aerial Images and a Comparison Study Remote Sensing building edge extraction high resolution image semantic segmentation edge detection deep learning |
title | Building Outline Extraction Directly Using the U<sup>2</sup>-Net Semantic Segmentation Model from High-Resolution Aerial Images and a Comparison Study |
title_full | Building Outline Extraction Directly Using the U<sup>2</sup>-Net Semantic Segmentation Model from High-Resolution Aerial Images and a Comparison Study |
title_fullStr | Building Outline Extraction Directly Using the U<sup>2</sup>-Net Semantic Segmentation Model from High-Resolution Aerial Images and a Comparison Study |
title_full_unstemmed | Building Outline Extraction Directly Using the U<sup>2</sup>-Net Semantic Segmentation Model from High-Resolution Aerial Images and a Comparison Study |
title_short | Building Outline Extraction Directly Using the U<sup>2</sup>-Net Semantic Segmentation Model from High-Resolution Aerial Images and a Comparison Study |
title_sort | building outline extraction directly using the u sup 2 sup net semantic segmentation model from high resolution aerial images and a comparison study |
topic | building edge extraction high resolution image semantic segmentation edge detection deep learning |
url | https://www.mdpi.com/2072-4292/13/16/3187 |
work_keys_str_mv | AT xinchunwei buildingoutlineextractiondirectlyusingtheusup2supnetsemanticsegmentationmodelfromhighresolutionaerialimagesandacomparisonstudy AT xingli buildingoutlineextractiondirectlyusingtheusup2supnetsemanticsegmentationmodelfromhighresolutionaerialimagesandacomparisonstudy AT weiliu buildingoutlineextractiondirectlyusingtheusup2supnetsemanticsegmentationmodelfromhighresolutionaerialimagesandacomparisonstudy AT lianpengzhang buildingoutlineextractiondirectlyusingtheusup2supnetsemanticsegmentationmodelfromhighresolutionaerialimagesandacomparisonstudy AT dayucheng buildingoutlineextractiondirectlyusingtheusup2supnetsemanticsegmentationmodelfromhighresolutionaerialimagesandacomparisonstudy AT hanyuji buildingoutlineextractiondirectlyusingtheusup2supnetsemanticsegmentationmodelfromhighresolutionaerialimagesandacomparisonstudy AT wenzhengzhang buildingoutlineextractiondirectlyusingtheusup2supnetsemanticsegmentationmodelfromhighresolutionaerialimagesandacomparisonstudy AT kaiyuan buildingoutlineextractiondirectlyusingtheusup2supnetsemanticsegmentationmodelfromhighresolutionaerialimagesandacomparisonstudy |