Scene classification for aerial images based on CNN using sparse coding technique

Abdul Qayyum, Aamir Saeed Malik, Naufal M. Saad, Mahboob Iqbal, Mohd Faris Abdullah, Waqas Rasheed, Tuan Ab Rashid Tuan Abdullah, Mohd Yaqoob Bin Jafaar

Research output: Contribution to journalArticle

19 Citations (Scopus)

Abstract

Aerial scene classification purposes to automatically label aerial images with specific semantic categories. However, cataloguing presents a fundamental problem for high-resolution remote-sensing imagery (HRRS). Recent developments include several approaches and numerous algorithms address the task. This article proposes a convolutional neural network (CNN) approach that utilizes sparse coding for scene classification applicable for HRRS unmanned aerial vehicle (UAV) and satellite imagery. The article has two major sections: the first describes the extraction of dense multiscale features (multiple scales) from the last convolutional layer of a pre-trained CNN models; the second describes the encoding of extracted features into global image features via sparse coding to achieve scene classification. The authors compared experimental outcomes with existing techniques such as Scale-Invariant Feature Transform and demonstrated that features from pre-trained CNNs generalized well with HRRS datasets and were more expressive than low- and mid-level features, exhibiting an overall 90.3% accuracy rate for scene classification compared to 85.4% achieved by SIFT with sparse coding. Thus, the proposed CNN-based sparse coding approach obtained a robust performance that holds promising potential for future applications in satellite and UAV imaging.

Original languageEnglish
Pages (from-to)2662-2685
Number of pages24
JournalInternational Journal of Remote Sensing
Volume38
Issue number8-10
DOIs
Publication statusPublished - 19 May 2017

Fingerprint

imagery
remote sensing
satellite imagery
transform
vehicle
rate

All Science Journal Classification (ASJC) codes

  • Earth and Planetary Sciences(all)

Cite this

Qayyum, A., Malik, A. S., Saad, N. M., Iqbal, M., Faris Abdullah, M., Rasheed, W., ... Bin Jafaar, M. Y. (2017). Scene classification for aerial images based on CNN using sparse coding technique. International Journal of Remote Sensing, 38(8-10), 2662-2685. https://doi.org/10.1080/01431161.2017.1296206
Qayyum, Abdul ; Malik, Aamir Saeed ; Saad, Naufal M. ; Iqbal, Mahboob ; Faris Abdullah, Mohd ; Rasheed, Waqas ; Tuan Abdullah, Tuan Ab Rashid ; Bin Jafaar, Mohd Yaqoob. / Scene classification for aerial images based on CNN using sparse coding technique. In: International Journal of Remote Sensing. 2017 ; Vol. 38, No. 8-10. pp. 2662-2685.
@article{b24d4da1188d4706bce83f66be497e73,
title = "Scene classification for aerial images based on CNN using sparse coding technique",
abstract = "Aerial scene classification purposes to automatically label aerial images with specific semantic categories. However, cataloguing presents a fundamental problem for high-resolution remote-sensing imagery (HRRS). Recent developments include several approaches and numerous algorithms address the task. This article proposes a convolutional neural network (CNN) approach that utilizes sparse coding for scene classification applicable for HRRS unmanned aerial vehicle (UAV) and satellite imagery. The article has two major sections: the first describes the extraction of dense multiscale features (multiple scales) from the last convolutional layer of a pre-trained CNN models; the second describes the encoding of extracted features into global image features via sparse coding to achieve scene classification. The authors compared experimental outcomes with existing techniques such as Scale-Invariant Feature Transform and demonstrated that features from pre-trained CNNs generalized well with HRRS datasets and were more expressive than low- and mid-level features, exhibiting an overall 90.3{\%} accuracy rate for scene classification compared to 85.4{\%} achieved by SIFT with sparse coding. Thus, the proposed CNN-based sparse coding approach obtained a robust performance that holds promising potential for future applications in satellite and UAV imaging.",
author = "Abdul Qayyum and Malik, {Aamir Saeed} and Saad, {Naufal M.} and Mahboob Iqbal and {Faris Abdullah}, Mohd and Waqas Rasheed and {Tuan Abdullah}, {Tuan Ab Rashid} and {Bin Jafaar}, {Mohd Yaqoob}",
year = "2017",
month = "5",
day = "19",
doi = "10.1080/01431161.2017.1296206",
language = "English",
volume = "38",
pages = "2662--2685",
journal = "International Joural of Remote Sensing",
issn = "0143-1161",
publisher = "Taylor and Francis Ltd.",
number = "8-10",

}

Qayyum, A, Malik, AS, Saad, NM, Iqbal, M, Faris Abdullah, M, Rasheed, W, Tuan Abdullah, TAR & Bin Jafaar, MY 2017, 'Scene classification for aerial images based on CNN using sparse coding technique', International Journal of Remote Sensing, vol. 38, no. 8-10, pp. 2662-2685. https://doi.org/10.1080/01431161.2017.1296206

Scene classification for aerial images based on CNN using sparse coding technique. / Qayyum, Abdul; Malik, Aamir Saeed; Saad, Naufal M.; Iqbal, Mahboob; Faris Abdullah, Mohd; Rasheed, Waqas; Tuan Abdullah, Tuan Ab Rashid; Bin Jafaar, Mohd Yaqoob.

In: International Journal of Remote Sensing, Vol. 38, No. 8-10, 19.05.2017, p. 2662-2685.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Scene classification for aerial images based on CNN using sparse coding technique

AU - Qayyum, Abdul

AU - Malik, Aamir Saeed

AU - Saad, Naufal M.

AU - Iqbal, Mahboob

AU - Faris Abdullah, Mohd

AU - Rasheed, Waqas

AU - Tuan Abdullah, Tuan Ab Rashid

AU - Bin Jafaar, Mohd Yaqoob

PY - 2017/5/19

Y1 - 2017/5/19

N2 - Aerial scene classification purposes to automatically label aerial images with specific semantic categories. However, cataloguing presents a fundamental problem for high-resolution remote-sensing imagery (HRRS). Recent developments include several approaches and numerous algorithms address the task. This article proposes a convolutional neural network (CNN) approach that utilizes sparse coding for scene classification applicable for HRRS unmanned aerial vehicle (UAV) and satellite imagery. The article has two major sections: the first describes the extraction of dense multiscale features (multiple scales) from the last convolutional layer of a pre-trained CNN models; the second describes the encoding of extracted features into global image features via sparse coding to achieve scene classification. The authors compared experimental outcomes with existing techniques such as Scale-Invariant Feature Transform and demonstrated that features from pre-trained CNNs generalized well with HRRS datasets and were more expressive than low- and mid-level features, exhibiting an overall 90.3% accuracy rate for scene classification compared to 85.4% achieved by SIFT with sparse coding. Thus, the proposed CNN-based sparse coding approach obtained a robust performance that holds promising potential for future applications in satellite and UAV imaging.

AB - Aerial scene classification purposes to automatically label aerial images with specific semantic categories. However, cataloguing presents a fundamental problem for high-resolution remote-sensing imagery (HRRS). Recent developments include several approaches and numerous algorithms address the task. This article proposes a convolutional neural network (CNN) approach that utilizes sparse coding for scene classification applicable for HRRS unmanned aerial vehicle (UAV) and satellite imagery. The article has two major sections: the first describes the extraction of dense multiscale features (multiple scales) from the last convolutional layer of a pre-trained CNN models; the second describes the encoding of extracted features into global image features via sparse coding to achieve scene classification. The authors compared experimental outcomes with existing techniques such as Scale-Invariant Feature Transform and demonstrated that features from pre-trained CNNs generalized well with HRRS datasets and were more expressive than low- and mid-level features, exhibiting an overall 90.3% accuracy rate for scene classification compared to 85.4% achieved by SIFT with sparse coding. Thus, the proposed CNN-based sparse coding approach obtained a robust performance that holds promising potential for future applications in satellite and UAV imaging.

UR - http://www.scopus.com/inward/record.url?scp=85014553885&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85014553885&partnerID=8YFLogxK

U2 - 10.1080/01431161.2017.1296206

DO - 10.1080/01431161.2017.1296206

M3 - Article

AN - SCOPUS:85014553885

VL - 38

SP - 2662

EP - 2685

JO - International Joural of Remote Sensing

JF - International Joural of Remote Sensing

SN - 0143-1161

IS - 8-10

ER -