Change Detection Based on Low-Level to High-Level Features Integration With Limited Samples
Detailed land cover change in multitemporal images is an important application for earth science. Many techniques have been proposed to solve this problem in different ways. However, accurately identifying changes still remains a challenge due to the difficulties in describing the characteristics of...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2020-01-01
|
Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9216597/ |
id |
doaj-5f3022ddeccf43e3a97785886cf8782b |
---|---|
record_format |
Article |
spelling |
doaj-5f3022ddeccf43e3a97785886cf8782b2021-06-03T23:03:35ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352020-01-01136260627610.1109/JSTARS.2020.30294609216597Change Detection Based on Low-Level to High-Level Features Integration With Limited SamplesXin Wang0https://orcid.org/0000-0001-6653-2585Peijun Du1https://orcid.org/0000-0002-2488-2656Dongmei Chen2https://orcid.org/0000-0001-5419-8735Sicong Liu3https://orcid.org/0000-0003-1612-4844Wei Zhang4https://orcid.org/0000-0001-8162-9422Erzhu Li5Jiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources, School of Geography and Ocean Science, Nanjing University, Nanjing, ChinaJiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources, School of Geography and Ocean Science, Nanjing University, Nanjing, ChinaDepartment of Geography and Planning, Queen's University, Kingston, ON, CanadaCollege of Surveying and Geoinformatics, Tongji University, Shanghai, ChinaJiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Key Laboratory for Land Satellite Remote Sensing Applications of Ministry of Natural Resources, School of Geography and Ocean Science, Nanjing University, Nanjing, ChinaSchool of Geography, Geomatics and Planning, Jiangsu Normal University, Xuzhou, ChinaDetailed land cover change in multitemporal images is an important application for earth science. Many techniques have been proposed to solve this problem in different ways. However, accurately identifying changes still remains a challenge due to the difficulties in describing the characteristics of various change categories by using single-level features. In this article, a multilevel feature representation framework was designed to build robust feature set for complex change detection task. First, four different levels of information from low level to high level, including pixel-level, neighborhood-level, object-level, and scene-level features, were extracted. Through the operation of extracting different level features from multitemporal images, the differences between them can be described comprehensively. Second, multilevel features were fused to reduce the dimension and then used as the input for supervised change detector with initial limited labels. Finally, for reducing the labeling cost and improving the change detection results simultaneously, active learning was conducted to select the most informative samples for labeling, and this step together with the previous steps were iteratively conducted to improve the results in each round. Experimental results of three pairs of real remote sensing datasets demonstrated that the proposed framework outperformed the other state-of-the-art methods in terms of accuracy. Moreover, the influences of scene scale for high-level semantic features in the proposed approach on change detection performance were also analyzed and discussed.https://ieeexplore.ieee.org/document/9216597/Active learningattribute profiles (APs)change detectionconvolutional neural network (CNN)multilevel featureobject feature |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Xin Wang Peijun Du Dongmei Chen Sicong Liu Wei Zhang Erzhu Li |
spellingShingle |
Xin Wang Peijun Du Dongmei Chen Sicong Liu Wei Zhang Erzhu Li Change Detection Based on Low-Level to High-Level Features Integration With Limited Samples IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Active learning attribute profiles (APs) change detection convolutional neural network (CNN) multilevel feature object feature |
author_facet |
Xin Wang Peijun Du Dongmei Chen Sicong Liu Wei Zhang Erzhu Li |
author_sort |
Xin Wang |
title |
Change Detection Based on Low-Level to High-Level Features Integration With Limited Samples |
title_short |
Change Detection Based on Low-Level to High-Level Features Integration With Limited Samples |
title_full |
Change Detection Based on Low-Level to High-Level Features Integration With Limited Samples |
title_fullStr |
Change Detection Based on Low-Level to High-Level Features Integration With Limited Samples |
title_full_unstemmed |
Change Detection Based on Low-Level to High-Level Features Integration With Limited Samples |
title_sort |
change detection based on low-level to high-level features integration with limited samples |
publisher |
IEEE |
series |
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
issn |
2151-1535 |
publishDate |
2020-01-01 |
description |
Detailed land cover change in multitemporal images is an important application for earth science. Many techniques have been proposed to solve this problem in different ways. However, accurately identifying changes still remains a challenge due to the difficulties in describing the characteristics of various change categories by using single-level features. In this article, a multilevel feature representation framework was designed to build robust feature set for complex change detection task. First, four different levels of information from low level to high level, including pixel-level, neighborhood-level, object-level, and scene-level features, were extracted. Through the operation of extracting different level features from multitemporal images, the differences between them can be described comprehensively. Second, multilevel features were fused to reduce the dimension and then used as the input for supervised change detector with initial limited labels. Finally, for reducing the labeling cost and improving the change detection results simultaneously, active learning was conducted to select the most informative samples for labeling, and this step together with the previous steps were iteratively conducted to improve the results in each round. Experimental results of three pairs of real remote sensing datasets demonstrated that the proposed framework outperformed the other state-of-the-art methods in terms of accuracy. Moreover, the influences of scene scale for high-level semantic features in the proposed approach on change detection performance were also analyzed and discussed. |
topic |
Active learning attribute profiles (APs) change detection convolutional neural network (CNN) multilevel feature object feature |
url |
https://ieeexplore.ieee.org/document/9216597/ |
work_keys_str_mv |
AT xinwang changedetectionbasedonlowleveltohighlevelfeaturesintegrationwithlimitedsamples AT peijundu changedetectionbasedonlowleveltohighlevelfeaturesintegrationwithlimitedsamples AT dongmeichen changedetectionbasedonlowleveltohighlevelfeaturesintegrationwithlimitedsamples AT sicongliu changedetectionbasedonlowleveltohighlevelfeaturesintegrationwithlimitedsamples AT weizhang changedetectionbasedonlowleveltohighlevelfeaturesintegrationwithlimitedsamples AT erzhuli changedetectionbasedonlowleveltohighlevelfeaturesintegrationwithlimitedsamples |
_version_ |
1721398673279025152 |