Ensemble Learning for Multispectral Scene Classification

被引:2
作者
Soroush, Rahman [1 ]
Baleghi, Yasser [1 ]
机构
[1] Babol Noshirvani Univ Technol Babol, Dept Elect & Comp Engn, Mazandaran, Iran
关键词
Convolutional neural networks; transfer learning; ensemble learning; infrared and visible images; scene recognition; CONVOLUTIONAL NEURAL-NETWORK; MODEL; CNN;
D O I
10.1142/S0218001422510132
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In the recent decades, various techniques based on deep convolutional neural networks (DCNNs) have been applied to scene classification. Most of the techniques are established upon single-spectral images such that environmental conditions may greatly affect the quality of images in the visible (RGB) spectrum. One remedy for this downside is to merge the infrared (IR) with the visible spectrum for gaining the complementary information in comparison with the unimodal analysis. This paper incorporates the RGB, IR and near-infrared (NIR) images into a multispectral analysis for scene classification. For this purpose, two strategies are adopted. In the first strategy, each RGB, IR and NIR image is separately applied to DCNNs and then classified according to the output score of each network. In addition, an optimal decision threshold is obtained based on the same output score of each network. In the second strategy, three image components are extracted from each type of image using wavelet transform decomposition. Independent DCNNs are then trained on the image components of all the scene classes. Eventually, the final classification of the scene is accomplished through an appropriate ensemble architecture. The use of this architecture alongside a transfer learning approach and simple classifiers leads to lesser computational costs in small datasets. These experiments reveal the superiority of the proposed method over the state-of-the-art architectures in terms of the accuracy of scene classification.
引用
收藏
页数:26
相关论文
共 53 条
[1]   Classification of Remote Sensing Images Using EfficientNet-B3 CNN Model With Attention [J].
Alhichri, Haikel ;
Alswayed, Asma S. ;
Bazi, Yakoub ;
Ammour, Nassim ;
Alajlan, Naif A. .
IEEE ACCESS, 2021, 9 :14078-14094
[2]   Ensemble of CNN for multi-focus image fusion [J].
Amin-Naji, Mostafa ;
Aghagolzadeh, Ali ;
Ezoji, Mehdi .
INFORMATION FUSION, 2019, 51 :201-214
[3]  
Asvadi A., 2012, RSF-RUS SAGE J SOC S, V4, P29
[4]   Softly combining an ensemble of classifiers learned from a single convolutional neural network for scene categorization [J].
Bai, Shuang ;
Tang, Huadong .
APPLIED SOFT COMPUTING, 2018, 67 :183-196
[5]   Scene classification using a hybrid generative/discriminative approach [J].
Bosch, Anna ;
Zisserman, Andrew ;
Munoz, Xavier .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2008, 30 (04) :712-727
[6]  
Bosch A, 2006, LECT NOTES COMPUT SC, V3954, P517
[7]  
Brown M, 2011, PROC CVPR IEEE, P177, DOI 10.1109/CVPR.2011.5995637
[8]   Adopting deep learning methods for airborne RGB fluvial scene classification [J].
Carbonneau, Patrice E. ;
Dugdale, Stephen J. ;
Breckon, Toby P. ;
Dietrich, James T. ;
Fonstad, Mark A. ;
Miyamoto, Hitoshi ;
Woodget, Amy S. .
REMOTE SENSING OF ENVIRONMENT, 2020, 251
[9]   Military Image Scene Recognition Based on CNN and Semantic Information [J].
Chen Cheng ;
Pan Chongyu ;
Huang Jian ;
Yuan Xingsheng .
2018 3RD INTERNATIONAL CONFERENCE ON MECHANICAL, CONTROL AND COMPUTER ENGINEERING (ICMCCE), 2018, :573-577
[10]   Scene recognition with objectness [J].
Cheng, Xiaojuan ;
Lu, Jiwen ;
Feng, Jianjiang ;
Yuan, Bo ;
Zhou, Jie .
PATTERN RECOGNITION, 2018, 74 :474-487