Building extraction method for aerial images based on DeepLabv3+ semantic segmentation

Aerial imagery can provide rich geographic information. As an important ground object information, quickly and accurately extracting buildings from aerial images can achieve target monitoring, location positioning, and further enrich specific geographic information in a given area. To address the is...

Full description

Saved in:
Bibliographic Details
Main Author: LIAO Yuanhui, WANG Jingdong, LI Haoran, YANG Heng
Format: Article
Language:zho
Published: Editorial Office of Command Control and Simulation 2024-12-01
Series:Zhihui kongzhi yu fangzhen
Subjects:
Online Access:https://www.zhkzyfz.cn/fileup/1673-3819/PDF/1732684200431-2040351917.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846151706715357184
author LIAO Yuanhui, WANG Jingdong, LI Haoran, YANG Heng
author_facet LIAO Yuanhui, WANG Jingdong, LI Haoran, YANG Heng
author_sort LIAO Yuanhui, WANG Jingdong, LI Haoran, YANG Heng
collection DOAJ
description Aerial imagery can provide rich geographic information. As an important ground object information, quickly and accurately extracting buildings from aerial images can achieve target monitoring, location positioning, and further enrich specific geographic information in a given area. To address the issues of segmentation result merging and irregular contour lines in semantic segmentation algorithms for building extraction, an improved model based on DeepLabv3+ for aerial building extraction is proposed by improving the feature fusion structure, constructing a comprehensive loss function, and incorporating an improved Douglas Peucker algorithm. Experimental results show that the improved model achieves an IoU of 0.794 on the test set, a 14.7% improvement compared to the original model. It effectively avoids the problem of merged segmentation between neighboring buildings and results in more regular segmentation boundaries, enabling more accurate extraction of the building contours.
format Article
id doaj-art-a30ad6ca7fbd4cb4aaacd86ddeaf31c9
institution Kabale University
issn 1673-3819
language zho
publishDate 2024-12-01
publisher Editorial Office of Command Control and Simulation
record_format Article
series Zhihui kongzhi yu fangzhen
spelling doaj-art-a30ad6ca7fbd4cb4aaacd86ddeaf31c92024-11-27T05:28:12ZzhoEditorial Office of Command Control and SimulationZhihui kongzhi yu fangzhen1673-38192024-12-01466556110.3969/j.issn.1673-3819.2024.06.010Building extraction method for aerial images based on DeepLabv3+ semantic segmentationLIAO Yuanhui, WANG Jingdong, LI Haoran, YANG Heng0College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, ChinaAerial imagery can provide rich geographic information. As an important ground object information, quickly and accurately extracting buildings from aerial images can achieve target monitoring, location positioning, and further enrich specific geographic information in a given area. To address the issues of segmentation result merging and irregular contour lines in semantic segmentation algorithms for building extraction, an improved model based on DeepLabv3+ for aerial building extraction is proposed by improving the feature fusion structure, constructing a comprehensive loss function, and incorporating an improved Douglas Peucker algorithm. Experimental results show that the improved model achieves an IoU of 0.794 on the test set, a 14.7% improvement compared to the original model. It effectively avoids the problem of merged segmentation between neighboring buildings and results in more regular segmentation boundaries, enabling more accurate extraction of the building contours.https://www.zhkzyfz.cn/fileup/1673-3819/PDF/1732684200431-2040351917.pdfaerial image|building extraction|deeplabv3+|contour regularization
spellingShingle LIAO Yuanhui, WANG Jingdong, LI Haoran, YANG Heng
Building extraction method for aerial images based on DeepLabv3+ semantic segmentation
Zhihui kongzhi yu fangzhen
aerial image|building extraction|deeplabv3+|contour regularization
title Building extraction method for aerial images based on DeepLabv3+ semantic segmentation
title_full Building extraction method for aerial images based on DeepLabv3+ semantic segmentation
title_fullStr Building extraction method for aerial images based on DeepLabv3+ semantic segmentation
title_full_unstemmed Building extraction method for aerial images based on DeepLabv3+ semantic segmentation
title_short Building extraction method for aerial images based on DeepLabv3+ semantic segmentation
title_sort building extraction method for aerial images based on deeplabv3 semantic segmentation
topic aerial image|building extraction|deeplabv3+|contour regularization
url https://www.zhkzyfz.cn/fileup/1673-3819/PDF/1732684200431-2040351917.pdf
work_keys_str_mv AT liaoyuanhuiwangjingdonglihaoranyangheng buildingextractionmethodforaerialimagesbasedondeeplabv3semanticsegmentation