基于空间域目标显著性分析的波段选择方法

doi:10.3964/j.issn.1000-0593(2023)09-2952-08

摘要
参考文献
相关文章 (15)

全文: PDF (39826 KB)
输出: BibTeX | EndNote (RIS)

摘要：作为遥感领域的新兴技术，高光谱成像为遥感影像处理分析和计算机视觉提供了海量内容。高光谱图像的优势在于电磁波谱的范围广度与高分辨率，能够将地物目标的光谱反射特性和差异特征更全面地表现出来，广泛地应用于地物分类、目标识别、异常检测等领域。但是，高光谱图像由于数据量繁重、信息重叠冗杂等问题，给图像处理、存储和传输带来一定挑战。选择合适的光谱波段可以在不改变原图像物理信息的情况下，达到较好的图像处理成果。为设计适合数据降维和目标地物分类的波段选择方法，提出将视觉显著性模型应用到波段选择方法中。首先引入基于图像空间分布的目标显著性算法进行波段图像处理得到目标显著图；其次，利用目标显著图分析地物之间在每一波段图像中的可分离程度定义为波段显著性。为避免波段信息重叠，在波段选择之前利用谱聚类算法将波段划分为若干子空间。然后在子空间内依据波段显著性降序排列，选择各子空间中目标显著性表现较好的波段组成波段子集；最后，在GF-5采集的高光谱图像数据进行方法验证，筛选有效的目标显著性算法，与常用的波段选择算法进行分类精度比较。结果表明，基于LC目标显著性算法的波段选择子集，在SVM分类器中具有优异分类结果，总体分类精度和Kappa系数达87.780 0%和0.805 3，优于应用全波段和其他三种波段选择方法的结果子集。

关键词：高光谱遥感；数据降维；目标显著性；波段选择；地物分类

Abstract：As an emerging technology in remote sensing, hyperspectral imaging provides massive content for remote sensing image processing analysis and computer vision. Hyperspectral images' advantages lie in the wide and high resolution of the electromagnetic spectrum, which can show the inherent spectral reflection characteristics of ground objects in a more comprehensive and discriminating manner, and are widely used in ground object classification, target recognition, anomaly detection, etc. However, its huge amount of data and redundant information causes considerable difficulties for hyperspectral image processing, storage and transmission. Band selection is a data dimensionality reduction method that can effectively reduce the amount of image data without changing the physical information of the original image. In order to achieve a better classification effect of ground objects, the visual saliency model is applied to the band selection method. Firstly, the target saliency algorithm based on image space distribution is introduced to process the band image to obtain the target saliency map. Secondly, using the target saliency map to analyze the degree of separability between ground objects in each band image is defined as band saliency. Spectral clustering algorithm is used to divide bands into several subspaces before band selection. Then in the subspace, the bands are sorted in descending order according to the saliency of the bands, and the bands with better target saliency performance in each subspace are selected to form the band subsets. Finally, the method is verified on the hyperspectral image data collected by GF-5, the effective target saliency algorithm is screened, and the classification accuracy is compared with the commonly used band selection algorithm. The experimental results show that the band selection subset based on the LC target saliency algorithm has excellent classification results in the SVM classifier, with overall classification accuracy and Kappa coefficient of 87.780 0% and 0.805 3. This method outperforms the results of the other three band selection methods and the results of all bands.

Key words：Hyperspectral remote sensing; Data reduction; Target saliency; Band selection; Object classification

收稿日期: 2022-05-26 修订日期: 2022-09-22

中图分类号:

TP751.1

基金资助: 国家自然科学基金项目（41971290）资助

通讯作者: 杨桄，卢珊 E-mail: yg2599@126.com; lus123@nenu.edu.cn

作者简介: 金椿柏，1997年生，空军航空大学硕士研究生 e-mail: 1262147682@qq.com

引用本文:

金椿柏，杨桄，卢珊，刘文婧，李德军，郑南. 基于空间域目标显著性分析的波段选择方法[J]. 光谱学与光谱分析, 2023, 43(09): 2952-2959.
JIN Chun-bai, YANG Guang, LU Shan, LIU Wen-jing, LI De-jun, ZHENG Nan. Band Selection Method Based on Target Saliency Analysis in Spatial Domain. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(09): 2952-2959.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2023)09-2952-08 或 https://www.gpxygpfx.com/CN/Y2023/V43/I09/2952

[1] Plaza A, Martinez P, Plaza J, et al. IEEE Transactions on Geoscience and Remote Sensing, 2005, 43(3): 466.
[2] Xie L, Li G, Peng L, et al. Journal of Applied Remote Sensing, 2017, 11(2): 026018.
[3] Sun H, Ren J, Zhao H, et al. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: doi: 10.1109/TGRS.2021.3075663.
[4] Wang Q, Li Q, Li X. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2019, 12(12): 4940.
[5] Cai Y, Liu X, Cai Z. IEEE Transactions on Geoscience and Remote Sensing, 2020, 58(3): 1969.
[6] Dou D, Gao K, Zhang X. IEEE Geoscience and Remote Sensing Letters, 2021，18(1)：147.
[7] Sun W, Peng J, Yang G, et al. IEEE Transactions on Geoscience and Remote Sensing, 2020, 58(6): 3906.
[8] Sui C, Zhou J, Li C, et al. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60：5504815.
[9] Zhao M, Yue L, Hu J, et al. IET Image Processing, 2021, 15(11): 2301.
[10] Yan C, Jing Z, Qi T, et al. Salient Target Detection in Hyperspectral Images Using Spectral Saliency. IEEE China Summit & International Conference on Signal and Information Processing, 2015.
[11] Yan H, Zhang Y, Wei W, et al. Salient Object Detection in Hyperspectral Imagery Using Spectral Gradient Contrast. IGARSS 2016 IEEE International Geoscience and Remote Sensing Symposium. IEEE, 2016, 1560.
[12] Moan S L, Mansouri A, Hardeberg J Y, et al. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013, 6(6): 2472.
[13] GONG Ran（龚燃）. Satellite Application（卫星应用）, 2018，(5): 10.3969/j.issn.1674-9030.2018.05.018.
[14] Hou X, Zhang L. Saliency Detection: A spectral Residual Approach, 2007 IEEE Conference on Computer Vision and Pattern Recognition, 2007.
[15] Achanta R, Hemami S S, Estrada F J, et al. Frequency-Tuned Salient Region Detection. IEEE Conference on Computer Vision and Pattern Recognition, 2009.
[16] Cheng M M, Mitra N J, Huang X, et al. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2013, 37(3): 569.
[17] Zhai Y, Shah M. Proceedings of the 14th ACM International Conference on Multimedia, MM 2006. 2006, 815 (https://doi.org/10.1145/1180639.1180824).
[18] Borji A, Cheng M M, Jiang H, et al. IEEE Transactions on Image Processing, 2015, 24(12): 5706.
[19] Guo B，Gunn S R, Damper R I, et al. IEEE Geoscience and Remote Sensing Letters, 2006, 3: 522.
[20] YAN Yang, HUA Wen-shen, LIU Xun, et al（严阳, 华文深, 刘恂, 等）. Optical Technique（光学技术）, 2018, 44(5): 7.
[21] WANG Li-guo, ZHAO Liang, SHI Yao（王立国, 赵亮, 石瑶）. CAAI Transactions on Intelligent Systems（智能系统学报）, 2018, 13(1): 131.