A Multiscale and Multitask Deep Learning Framework for Automatic Building Extraction
Detecting buildings, segmenting building footprints, and extracting building edges from high-resolution remote sensing images are vital in applications such as urban planning, change detection, smart cities, and map-making and updating. The tasks of building detection, footprint segmentation, and ed...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-09-01
|
Series: | Remote Sensing |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-4292/14/19/4744 |
_version_ | 1797477198798520320 |
---|---|
author | Jichong Yin Fang Wu Yue Qiu Anping Li Chengyi Liu Xianyong Gong |
author_facet | Jichong Yin Fang Wu Yue Qiu Anping Li Chengyi Liu Xianyong Gong |
author_sort | Jichong Yin |
collection | DOAJ |
description | Detecting buildings, segmenting building footprints, and extracting building edges from high-resolution remote sensing images are vital in applications such as urban planning, change detection, smart cities, and map-making and updating. The tasks of building detection, footprint segmentation, and edge extraction affect each other to a certain extent. However, most previous works have focused on one of these three tasks and have lacked a multitask learning framework that can simultaneously solve the tasks of building detection, footprint segmentation and edge extraction, making it difficult to obtain smooth and complete buildings. This study proposes a novel multiscale and multitask deep learning framework to consider the dependencies among building detection, footprint segmentation, and edge extraction while completing all three tasks. In addition, a multitask feature fusion module is introduced into the deep learning framework to increase the robustness of feature extraction. A multitask loss function is also introduced to balance the training losses among the various tasks to obtain the best training results. Finally, the proposed method is applied to open-source building datasets and large-scale high-resolution remote sensing images and compared with other advanced building extraction methods. To verify the effectiveness of multitask learning, the performance of multitask learning and single-task training is compared in ablation experiments. The experimental results show that the proposed method has certain advantages over other methods and that multitask learning can effectively improve single-task performance. |
first_indexed | 2024-03-09T21:14:15Z |
format | Article |
id | doaj.art-b11b3c14953e415988189c194645992d |
institution | Directory Open Access Journal |
issn | 2072-4292 |
language | English |
last_indexed | 2024-03-09T21:14:15Z |
publishDate | 2022-09-01 |
publisher | MDPI AG |
record_format | Article |
series | Remote Sensing |
spelling | doaj.art-b11b3c14953e415988189c194645992d2023-11-23T21:37:55ZengMDPI AGRemote Sensing2072-42922022-09-011419474410.3390/rs14194744A Multiscale and Multitask Deep Learning Framework for Automatic Building ExtractionJichong Yin0Fang Wu1Yue Qiu2Anping Li3Chengyi Liu4Xianyong Gong5Institute of Geospatial Information, PLA Strategic Support Force Information Engineering University, Zhengzhou 450001, ChinaInstitute of Geospatial Information, PLA Strategic Support Force Information Engineering University, Zhengzhou 450001, ChinaInstitute of Geospatial Information, PLA Strategic Support Force Information Engineering University, Zhengzhou 450001, China78098 Troops, Chengdu 610000, ChinaInstitute of Geospatial Information, PLA Strategic Support Force Information Engineering University, Zhengzhou 450001, ChinaInstitute of Geospatial Information, PLA Strategic Support Force Information Engineering University, Zhengzhou 450001, ChinaDetecting buildings, segmenting building footprints, and extracting building edges from high-resolution remote sensing images are vital in applications such as urban planning, change detection, smart cities, and map-making and updating. The tasks of building detection, footprint segmentation, and edge extraction affect each other to a certain extent. However, most previous works have focused on one of these three tasks and have lacked a multitask learning framework that can simultaneously solve the tasks of building detection, footprint segmentation and edge extraction, making it difficult to obtain smooth and complete buildings. This study proposes a novel multiscale and multitask deep learning framework to consider the dependencies among building detection, footprint segmentation, and edge extraction while completing all three tasks. In addition, a multitask feature fusion module is introduced into the deep learning framework to increase the robustness of feature extraction. A multitask loss function is also introduced to balance the training losses among the various tasks to obtain the best training results. Finally, the proposed method is applied to open-source building datasets and large-scale high-resolution remote sensing images and compared with other advanced building extraction methods. To verify the effectiveness of multitask learning, the performance of multitask learning and single-task training is compared in ablation experiments. The experimental results show that the proposed method has certain advantages over other methods and that multitask learning can effectively improve single-task performance.https://www.mdpi.com/2072-4292/14/19/4744building extractionmultilevel and multitaskdeep learning frameworkhigh-resolution remote sensing image |
spellingShingle | Jichong Yin Fang Wu Yue Qiu Anping Li Chengyi Liu Xianyong Gong A Multiscale and Multitask Deep Learning Framework for Automatic Building Extraction Remote Sensing building extraction multilevel and multitask deep learning framework high-resolution remote sensing image |
title | A Multiscale and Multitask Deep Learning Framework for Automatic Building Extraction |
title_full | A Multiscale and Multitask Deep Learning Framework for Automatic Building Extraction |
title_fullStr | A Multiscale and Multitask Deep Learning Framework for Automatic Building Extraction |
title_full_unstemmed | A Multiscale and Multitask Deep Learning Framework for Automatic Building Extraction |
title_short | A Multiscale and Multitask Deep Learning Framework for Automatic Building Extraction |
title_sort | multiscale and multitask deep learning framework for automatic building extraction |
topic | building extraction multilevel and multitask deep learning framework high-resolution remote sensing image |
url | https://www.mdpi.com/2072-4292/14/19/4744 |
work_keys_str_mv | AT jichongyin amultiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT fangwu amultiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT yueqiu amultiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT anpingli amultiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT chengyiliu amultiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT xianyonggong amultiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT jichongyin multiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT fangwu multiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT yueqiu multiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT anpingli multiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT chengyiliu multiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction AT xianyonggong multiscaleandmultitaskdeeplearningframeworkforautomaticbuildingextraction |