光谱学与光谱分析 |
|
|
|
|
|
Clustering of Hyperspectral Image Based on Spatial-Spectral Chinese Restaurant Process Mixture Model |
SHU Yang1, LI Jing1, 2*, HE Shi1, TANG Hong1, 2, WANG Na1, 2, SHEN Li3, DU Hong-yue4 |
1. State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Normal University, Beijing 100875, China 2. Key Laboratory of Environmental Change and Natural Disaster, Ministry of Education, Beijing Normal University, Beijing 100875, China 3. Faculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu 611756, China 4. China Mapping Technology Service Corporation, Beijing 100088, China |
|
|
Abstract The classification of hyperspectral images is one of the most important study fields. The spectral information is used in traditional classification of hyperspectral images, while the spatial correlativity information is ignored. To solve this problem, a novel model called spatial-spectral Chinese restaurant process (ssCRP) is proposed to cluster the hyperspectral images, which is an extension of Chinese restaurant process. Both the spatial and spectral information are considered in the modeling and inference of the method. The proposed model clusters the hyperspectral images better than tradional methods and satisfies the requirement of hyperspectral image clustering. Firstly, in order to consider both spatial and spectral information, a new similarity measurement is defined withthe exponential decay function based on the spatial distance and spectral angle among pixels. Then, each pixel is associated with a table based on the table construction by considering the similarity. Finally, each table is allocated with a dish which corresponds to a cluster. Thus, each pixel of the hyperspectral image is allocated with a clustering label. The true hyperspectral image collected by airborne visible infrared imaging spectrometer (AVIRIS) is used to evaluate the performance of our model. Experimental results indicate that the proposed model outperforms traditional K-means and ISODATA. Compared with those of the two methods, the result of the proposed model is more regular with lower salt-and-pepper effect with higher spatial consistency. The classification accuracy of the proposed model reaches to 63.57% and the Kappa coefficient is 0.632 3, much higher than those of K-means and ISODATA. Meanwhile, the edges of the result of our model are well preserved.
|
Received: 2014-12-13
Accepted: 2015-03-19
|
|
Corresponding Authors:
LI Jing
E-mail: lijing@bnu.edu.cn
|
|
[1] Dale L M, Thewis A, Boudry C, et al. Applied Spectroscopy Reviews, 2013, 48(2): 142. [2] Ghiyamat A, Shafri H Z M. International Journal of Remote Sensing, 2010, 31(7): 1837. [3] Bazi Y, Melgani F. IEEE Transactions On Geoscience and Remote Sensing, 2006, 44(11): 3374. [4] Lee S, Crawford M M. IEEE International Geoscience and Remote Sensing Symposium, IGARSS’04,2004,2:941. [5] Tarabalka Y, Benediktsson J A, Chanussot J. IEEE Transactions on Geoscience And Remote Sensing, 2009, 47(8): 2973. [6] Alhichri H, Ammour N, Alajlan N, et al. Arabian Journal For Science and Engineering, 2014, 39(6): 3747. [7] LI Na, LI Yong-jie, ZHAO Hui-jie, et al(李 娜, 李咏洁, 赵慧洁, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2014, 34(2): 526. [8] Bouguila N, Ziou D. IEEE Transactions On Neural Networks, 2010, 21(1): 107. [9] Blei D M, Frazier P I. Journal of Machine Learning Research, 2011, 12: 2461. [10] Ghosh S, Ungureanu A B, Sudderth E B, et al. Advances in Neural Information Processing Systems, 2011. 1476. [11] Ferguson T S. The Annals of Statistics, 1973, 1(2): 209. [12] Teh Y W, Jordan M I, Beal M J, et al. Journal of the American Statistical Association, 2006, 101(476): 1566. [13] ZHOU Jian-ying, WANG Fei-yue, ZENG Da-jun(周建英, 王飞跃, 曾大军). Acta Automatica Sinica(自动化学报), 2011, 37(4): 389. [14] Frohn R C, Hao Y P. Remote Sensing of Environment, 2006, 100(2): 237. [15] Akcay H G, Aksoy S. IEEE Transactions on Geoscience And Remote Sensing, 2008, 46(7): 2097. |
[1] |
LIU Zhen1*, LIU Li2*, FAN Shuo2, ZHAO An-ran2, LIU Si-lu2. Training Sample Selection for Spectral Reconstruction Based on Improved K-Means Clustering[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2024, 44(01): 29-35. |
[2] |
GUO Ya-fei1, CAO Qiang1, YE Lei-lei1, ZHANG Cheng-yuan1, KOU Ren-bo1, WANG Jun-mei1, GUO Mei1, 2*. Double Index Sequence Analysis of FTIR and Anti-Inflammatory Spectrum Effect Relationship of Rheum Tanguticum[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2024, 44(01): 188-196. |
[3] |
XU Rong1, AO Dong-mei2*, LI Man-tian1, 2, LIU Sai1, GUO Kun1, HU Ying2, YANG Chun-mei2, XU Chang-qing1. Study on Traditional Chinese Medicine of Lonicera L. Based on Infrared Spectroscopy and Cluster Analysis[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(11): 3518-3523. |
[4] |
JIA Hao1, 3, 4, ZHANG Wei-fang1, 3, LEI Jing-wei1, 3*, LI Ying-ying1, 3, YANG Chun-jing2, 3*, XIE Cai-xia1, 3, GONG Hai-yan1, 3, DING Xin-yu1, YAO Tian-yi1. Study on Infrared Fingerprint of the Classical Famous
Prescription Yiguanjian[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(10): 3202-3210. |
[5] |
LI Xin-li1, CONG Li-li2, XU Shu-ping2, LI Su-yi1*. Cell Growth Analysis Method Based on Spectral Clustering and Single-Cell Raman Spectroscopy[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(09): 2832-2836. |
[6] |
FENG Hai-kuan1, 2, YUE Ji-bo3, FAN Yi-guang2, YANG Gui-jun2, ZHAO Chun-jiang1, 2*. Estimation of Potato Above-Ground Biomass Based on VGC-AGB Model and Hyperspectral Remote Sensing[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(09): 2876-2884. |
[7] |
JIN Chun-bai1, YANG Guang1*, LU Shan2*, LIU Wen-jing1, LI De-jun1, ZHENG Nan1. Band Selection Method Based on Target Saliency Analysis in Spatial Domain[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(09): 2952-2959. |
[8] |
GAO Yu1, SUN Xue-jian1*, LI Guang-hua2, ZHANG Li-fu1, QU Liang2, ZHANG Dong-hui1, CHANG Jing-jing2, DAI Xiao-ai3. Study on the Derivation of Paper Viscosity Spectral Index Based on Spectral Information Expansion[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(09): 2960-2966. |
[9] |
KONG Bo1, YU Huan2*, SONG Wu-jie2, 3, HOU Yu-ting2, XIANG Qing2. Hyperspectral Characteristics and Quantitative Remote Sensing Inversion of Gravel Grain Size in the North Tibetan Plateau[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(08): 2381-2390. |
[10] |
WANG Hui-min1, 2, YU Lei1, XU Kai-lei1, 2, JIANG Xiao-guang1, 2, WAN Yu-qing1, 2*. Estimation of Salt Content of Saline Soil in Arid Areas Based on GF-5 Hyperspectral Image[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(07): 2278-2286. |
[11] |
ZHANG Xia1, WANG Wei-hao1, 2*, SUN Wei-chao1, DING Song-tao1, 2, WANG Yi-bo1, 2. Soil Zn Content Inversion by Hyperspectral Remote Sensing Data and Considering Soil Types[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(07): 2019-2026. |
[12] |
CAO Yang1, 2, LI Yan-hong1, 2*. Study on the Effects of NO2 Pollution Under COVID-19 Epidemic
Prevention and Control in Urumqi[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(06): 1981-1987. |
[13] |
ZHANG Chao1*, SU Xiao-yu1, XIA Tian2, YANG Ke-ming3, FENG Fei-sheng4. Monitoring the Degree of Pollution in Different Varieties of Maize Under Copper and Lead Stress[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(04): 1268-1274. |
[14] |
LIU Si-qi1, FENG Guo-hong1*, TANG Jie2, REN Jia-qi1. Research on Identification of Wood Species by Mid-Infrared Spectroscopy Based on CA-SDP-DenseNet[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(03): 814-822. |
[15] |
LI Zi-yi1, LI Rui-lan1, LI Can-lin1, WANG Ke-ru2, FAN Jiu-yu3, GU Rui1*. Identification of Tibetan Medicine Zhaxun by Infrared Spectroscopy
Combined With Chemometrics[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(02): 526-532. |
|
|
|
|