Education, Science, Technology, Innovation and Life
Open Access
Sign In

Feature dimension reduction method of multi-source remote sensing image based on spatial information

Download as PDF

DOI: 10.23977/tranaa.2020.010101 | Downloads: 0 | Views: 19

Author(s)

Yulong Zhao 1, Min Hu 1

Affiliation(s)

1 Space Engineering University, Beijing, 101416, China

Corresponding Author

Yulong Zhao

ABSTRACT

The dimension reduction method of multi-source remote sensing image based on spatial information is proposed because of the influence of shadow on band selection and feature extraction. According to the separability of feature parameters, candidate feature parameters are selected to determine the classification threshold between samples. The feature parameters with separability greater than 6 are selected as the standard basis to construct the feature space of multi-source remote sensing image, so as to collect the feature space information of multi-source remote sensing image. According to the process of shadow detection algorithm, the original shadow image is transformed into HSV space to separate the dark area, and the dark area is denoised to get the shadow area. After shadow preprocessing, removing and post-processing, the boundary line is weakened and the undistorted information is obtained. According to the data structure of the multi-source remote sensing image, the image of each band in the multi-source remote sensing image can be expanded into one-dimensional vector by using the collected spatial information, and the independent components can be obtained by further solving. The independent component analysis (ICA) mathematical model was constructed to centralize the original mixed data and eliminate the mean value. The multi-source remote sensing image is whitened to form a new low-dimensional image. The experimental results show that the method is effective in dimensionality reduction.

KEYWORDS

Spatial information; Multi-source remote sensing image; Feature dimensionality reduction; Shadow detection and removal

CITE THIS PAPER

Yulong Zhao, Min Hu, Feature dimension reduction method of multi-source remote sensing image based on spatial information. Transactions on Aeronautics and Astronautics (2020) 1: 1-14. DOI: http://dx.doi.org/10.23977/tranaa.2020.010101.

REFERENCES

[1] Huang Hong, Shi Guangyao, Duan Yule, et al. Dimensionality reduction method of hyperspectral remote sensing image based on weighted space-spectrum joint preservation and embedding[J]. Acta Geomatica Sinica, 2019, 11(8):1014-1024. 
[2] Lu Jianqiang, Li Wangzhi, Lan Yubin, et al. Agricultural aerial remote sensing image registration algorithm based on point feature detection[J]. Transactions of the Chinese Society of Agricultural Engineering, 2020, 36(3): 71-77.
[3] Liu Yongmei, Ma Xiao, Men Chaoguang. Hyperspectral remote sensing image classification method based on multiple spatial information[J]. Chinese Space Science and Technology, 2019, 039(002): 73-81.
[4] Zhang Yan, Sun Jie, Yu Changchun, et al. Study on Quaternary Coverage Classification Method Based on Multi-source Remote Sensing Data: A Case Study of 1:50,000 Mapping of Qigandianzi Sheet in Inner Mongolia[J]. Geological Science and Technology Information , 2019, 38(02):287-296.
[5] Cheng Bo, Liu Yueming, Liu Xunan, et al. Research on high-resolution remote sensing image offshore aquaculture area extraction method based on multi-source feature fusion[J]. Remote Sensing Technology and Application, 2018, 33(2):296-304.
[6] Miao Kaichao, Luo Xichang, Zhang Shujing, et al. Feature extraction and grade recognition method of fog image based on color gamut analysis[J]. Science Technology and Engineering, 2019, 019(035):228-233.
[7] Tian Qinglin, Pan Wei, Li Hanbo, et al. Comparative study of alteration information extraction based on multi-source remote sensing data[J]. Geological Science and Technology Information, 2018, 37(06):224-231+244.
[8] Li Yunfeng, Gan Le, Lin Cong, et al. An urban rice field distribution extraction method based on multi-source and multi-temporal remote sensing information[J]. Geography and Geo-Information Science, 2018, v.34(03):53- 59.
[9] Long Yuhang, Wu Desheng. Three-dimensional virtual reconstruction simulation of spatial feature information of high-altitude remote sensing images[J]. Computer Simulation, 2019, 036(012):57-61.
[10] Meng Qingyan, Sun Yunxiao, Zhang Jiahui, et al. Analysis of the vertical distribution characteristics and spatial configuration of greening in the building area based on multi-source remote sensing——A case study of Szekszbek, Hungary[J]. Remote Sensing Technology and Application, 2018, 33 (002):370-376.

All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.