skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Domain-Adapted Convolutional Networks for Satellite Image Classification: A Large-Scale Interactive Learning Workflow

Abstract

Satellite imagery often exhibits large spatial extent areas that encompass object classes with considerable variability. This often limits large-scale model generalization with machine learning algorithms. Notably, acquisition conditions, including dates, sensor position, lighting condition, and sensor types, often translate into class distribution shifts introducing complex nonlinear factors and hamper the potential impact of machine learning classifiers. Here, this article investigates the challenge of exploiting satellite images using convolutional neural networks (CNN) for settlement classification where the class distribution shifts are significant. We present a large-scale human settlement mapping workflow based-off multiple modules to adapt a pretrained CNN to address the negative impact of distribution shift on classification performance. To extend a locally trained classifier onto large spatial extents areas we introduce several submodules: First, a human-in-the-loop element for relabeling of misclassified target domain samples to generate representative examples for model adaptation; second, an efficient hashing module to minimize redundancy and noisy samples from the mass-selected examples; and third, a novel relevance ranking module to minimize the dominance of source example on the target domain. The workflow presents a novel and practical approach to achieve large-scale domain adaptation with binary classifiers that are based-off CNN features. Experimental evaluations are conducted onmore » areas of interest that encompass various image characteristics, including multisensors, multitemporal, and multiangular conditions. Domain adaptation is assessed on source–target pairs through the transfer loss and transfer ratio metrics to illustrate the utility of the workflow.« less

Authors:
ORCiD logo [1]; ORCiD logo [1];  [1]; ORCiD logo [1]; ORCiD logo [1]; ORCiD logo [1]
  1. Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States). Computing and Computational Sciences Directorate
Publication Date:
Research Org.:
Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
1426590
Grant/Contract Number:  
AC05-00OR22725
Resource Type:
Journal Article: Accepted Manuscript
Journal Name:
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Additional Journal Information:
Journal Volume: 11; Journal Issue: 3; Journal ID: ISSN 1939-1404
Publisher:
IEEE
Country of Publication:
United States
Language:
English
Subject:
47 OTHER INSTRUMENTATION; 97 MATHEMATICS AND COMPUTING; Adaptation model; image classification; remote sensing; semi-supervised learning; supervised learning

Citation Formats

Lunga, Dalton D., Yang, Hsiuhan Lexie, Reith, Andrew E., Weaver, Jeanette E., Yuan, Jiangye, and Bhaduri, Budhendra L.. Domain-Adapted Convolutional Networks for Satellite Image Classification: A Large-Scale Interactive Learning Workflow. United States: N. p., 2018. Web. doi:10.1109/JSTARS.2018.2795753.
Lunga, Dalton D., Yang, Hsiuhan Lexie, Reith, Andrew E., Weaver, Jeanette E., Yuan, Jiangye, & Bhaduri, Budhendra L.. Domain-Adapted Convolutional Networks for Satellite Image Classification: A Large-Scale Interactive Learning Workflow. United States. doi:10.1109/JSTARS.2018.2795753.
Lunga, Dalton D., Yang, Hsiuhan Lexie, Reith, Andrew E., Weaver, Jeanette E., Yuan, Jiangye, and Bhaduri, Budhendra L.. Tue . "Domain-Adapted Convolutional Networks for Satellite Image Classification: A Large-Scale Interactive Learning Workflow". United States. doi:10.1109/JSTARS.2018.2795753. https://www.osti.gov/servlets/purl/1426590.
@article{osti_1426590,
title = {Domain-Adapted Convolutional Networks for Satellite Image Classification: A Large-Scale Interactive Learning Workflow},
author = {Lunga, Dalton D. and Yang, Hsiuhan Lexie and Reith, Andrew E. and Weaver, Jeanette E. and Yuan, Jiangye and Bhaduri, Budhendra L.},
abstractNote = {Satellite imagery often exhibits large spatial extent areas that encompass object classes with considerable variability. This often limits large-scale model generalization with machine learning algorithms. Notably, acquisition conditions, including dates, sensor position, lighting condition, and sensor types, often translate into class distribution shifts introducing complex nonlinear factors and hamper the potential impact of machine learning classifiers. Here, this article investigates the challenge of exploiting satellite images using convolutional neural networks (CNN) for settlement classification where the class distribution shifts are significant. We present a large-scale human settlement mapping workflow based-off multiple modules to adapt a pretrained CNN to address the negative impact of distribution shift on classification performance. To extend a locally trained classifier onto large spatial extents areas we introduce several submodules: First, a human-in-the-loop element for relabeling of misclassified target domain samples to generate representative examples for model adaptation; second, an efficient hashing module to minimize redundancy and noisy samples from the mass-selected examples; and third, a novel relevance ranking module to minimize the dominance of source example on the target domain. The workflow presents a novel and practical approach to achieve large-scale domain adaptation with binary classifiers that are based-off CNN features. Experimental evaluations are conducted on areas of interest that encompass various image characteristics, including multisensors, multitemporal, and multiangular conditions. Domain adaptation is assessed on source–target pairs through the transfer loss and transfer ratio metrics to illustrate the utility of the workflow.},
doi = {10.1109/JSTARS.2018.2795753},
journal = {IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing},
issn = {1939-1404},
number = 3,
volume = 11,
place = {United States},
year = {2018},
month = {2}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record

Citation Metrics:
Cited by: 1 work
Citation information provided by
Web of Science

Save / Share: