Scene classification based on a hierarchical convolutional sparse auto-encoder for high spatial resolution imagery
- 11 December 2016
- journal article
- research article
- Published by Taylor & Francis Ltd in International Journal of Remote Sensing
- Vol. 38 (2), 514-536
- https://doi.org/10.1080/01431161.2016.1266059
Abstract
Efficiently representing and recognizing the semantic classes of the subregions of large-scale high spatial resolution (HSR) remote-sensing images are challenging and critical problems. Most of the existing scene classification methods concentrate on the feature coding approach with handcrafted low-level features or the low-level unsupervised feature learning approaches, which essentially prevent them from better recognizing the semantic categories of the scene due to their limited mid-level feature representation ability. In this article, to overcome the inadequate mid-level representation, a patch-based spatial-spectral hierarchical convolutional sparse auto-encoder (HCSAE) algorithm, based on deep learning, is proposed for HSR remote-sensing imagery scene classification. The HCSAE framework uses an unsupervised hierarchical network based on a sparse auto-encoder (SAE) model. In contrast to the single-level SAE, the HCSAE framework utilizes the significant features from the single-level algorithm in a feedforward and full connection approach to the maximum extent, which adequately represents the scene semantics in the high level of the HCSAE. To ensure robust feature learning and extraction during the SAE feature extraction procedure, a ‘dropout’ strategy is also introduced. The experimental results using the UC Merced data set with 21 classes and a Google Earth data set with 12 classes demonstrate that the proposed HCSAE framework can provide better accuracy than the traditional scene classification methods and the single-level convolutional sparse auto-encoder (CSAE) algorithm.Keywords
Funding Information
- National Natural Science Foundation of China (41622107 and 41371344)
- Natural Science Foundation of Hubei Province (2016-29)
- State Key Laboratory of Earth Surface Processes and Resource Ecology (2015-KF-02)
This publication has 31 references indexed in Scilit:
- Scene classification via latent Dirichlet allocation using a hybrid generative/discriminative strategy for high spatial resolution remote sensing imageryRemote Sensing Letters, 2013
- Representation Learning: A Review and New PerspectivesIEEE Transactions on Pattern Analysis and Machine Intelligence, 2013
- Stacked Autoencoders for Unsupervised Feature Learning and Multiple Organ Detection in a Pilot Study Using 4D Patient DataIEEE Transactions on Pattern Analysis and Machine Intelligence, 2013
- Automatic landslide detection from remote-sensing imagery using a scene classification method based on BoVW and pLSAInternational Journal of Remote Sensing, 2012
- Per-pixel vs. object-based classification of urban land cover extraction using high spatial resolution imageryRemote Sensing of Environment, 2011
- On the classification of remote sensing high spatial resolution image dataInternational Journal of Remote Sensing, 2010
- Learning Deep Architectures for AIFoundations and Trends® in Machine Learning, 2009
- Reducing the Dimensionality of Data with Neural NetworksScience, 2006
- 10.1162/jmlr.2003.3.4-5.993Applied Physics Letters, 2000
- On the limited memory BFGS method for large scale optimizationMathematical Programming, 1989