Global Descriptors of Convolution Neural
Networks for Remote Scene Images
Classification
Q. Wang
1 , Qian Ning
1,2(B) , X. Yang
1 , Bingcai Chen
3,4 , Yinjie Lei
1 , C. Zhao
1 ,
T. Tang
1,2,3,4 , and R. Hu
1,2,3,4
1 College of Electrical & Information Engineering,
Sichuan University, Chengdu 610065, China
ningq@scu.edu.cn
2 School of Physics & Electronics,
Xinjiang Normal University, Urumqi 830054, China
3 School of Computer Science & Technology at Dalian University
of Technology, Dalian 116024, China
4 School of Computer Science & Technology at Xinjiang Normal University,
Urumqi 830054, China
Abstract. Nowadays, the deep learning-based methods have been
widely used in the scene-level-based image classification. However, the
features automatically obtained from the last fully connected (FC) layer
of single CNN without any process have little effect because of high
dimensionality. In this paper, we propose a simple enhancing scenelevel feature description method for remote sensing scene classification. Firstly, the principal component analysis (PCA) transformation is
adopted in our research for reducing redundant dimensionality. Secondly,
a new method is used to fuse features obtained by PCA transformation.
Finally, the random forest classifier applying to classification makes a
significant effect on compressing the training procedure. The results of
experiments on the public dataset describe that feature fusion with PCA
transformation performs great classification effect. Moreover, compared
with the classifier softmax, the random forest classifier outperforms the
softmax classifier in the training procedure.
Keywords: Remote sensing image (RSI) · Global feature descriptors ·
Feature fusion · Scene classification
1 Introduction
More recently, in the field of RSI investigation, RSI scene classification [1] is one
of the most important processes. For RSI classification, image semantic understanding is generally reflected by feature descriptors. Therefore, the key to classifying is features. In our research, we focus on the investigation of RSI feature
c
Springer Nature Singapore Pte Ltd. 2020
Q. Liang et al. (Eds.): Artificial Intelligence in China, LNEE 572, pp. 1–11, 2020.
https://doi.org/10.1007/978-981-15-0187-6_1
Networks for Remote Scene Images
Classification
Q. Wang
1 , Qian Ning
1,2(B) , X. Yang
1 , Bingcai Chen
3,4 , Yinjie Lei
1 , C. Zhao
1 ,
T. Tang
1,2,3,4 , and R. Hu
1,2,3,4
1 College of Electrical & Information Engineering,
Sichuan University, Chengdu 610065, China
ningq@scu.edu.cn
2 School of Physics & Electronics,
Xinjiang Normal University, Urumqi 830054, China
3 School of Computer Science & Technology at Dalian University
of Technology, Dalian 116024, China
4 School of Computer Science & Technology at Xinjiang Normal University,
Urumqi 830054, China
Abstract. Nowadays, the deep learning-based methods have been
widely used in the scene-level-based image classification. However, the
features automatically obtained from the last fully connected (FC) layer
of single CNN without any process have little effect because of high
dimensionality. In this paper, we propose a simple enhancing scenelevel feature description method for remote sensing scene classification. Firstly, the principal component analysis (PCA) transformation is
adopted in our research for reducing redundant dimensionality. Secondly,
a new method is used to fuse features obtained by PCA transformation.
Finally, the random forest classifier applying to classification makes a
significant effect on compressing the training procedure. The results of
experiments on the public dataset describe that feature fusion with PCA
transformation performs great classification effect. Moreover, compared
with the classifier softmax, the random forest classifier outperforms the
softmax classifier in the training procedure.
Keywords: Remote sensing image (RSI) · Global feature descriptors ·
Feature fusion · Scene classification
1 Introduction
More recently, in the field of RSI investigation, RSI scene classification [1] is one
of the most important processes. For RSI classification, image semantic understanding is generally reflected by feature descriptors. Therefore, the key to classifying is features. In our research, we focus on the investigation of RSI feature
c
Springer Nature Singapore Pte Ltd. 2020
Q. Liang et al. (Eds.): Artificial Intelligence in China, LNEE 572, pp. 1–11, 2020.
https://doi.org/10.1007/978-981-15-0187-6_1