Hyperspectral image classification via recursive filtering and KNN
Bing TU1,2,3, Xiaofei ZHANG1,3, Guoyun ZHANG1,2,3, Jinping WANG1,3, Yao ZHOU1,3
1.School of Information and Communication Engineering, Hunan Institute of Science and Technology, Yueyang 414006,China 2.Key Laboratory of Optimization and Control for Complex Systems of Hunan Province, Hunan Institute of Science and Technology, Yueyang 414006, China 3.Laboratory of Intelligent-Image Information Processing, Hunan Institute of Science and Technology, Yueyang 414006, China
In order to remove the noise in the hyperspectral image effectively, strengthen the spatial structure, make full use of the spatial context information of the object, and improve the classification accuracy of hyperspectral image, the authors put forward recursive filtering and k-nearest neighbor (KNN) method for hyperspectral image classification. The main steps are as follows: Firstly, the principal component analysis (PCA) is used to perform feature dimension reduction of hyperspectral images. Next, the recursive filtering is used to filter the principal component image. Then, the Euclidean distance between the test sample and the different training samples is calculated by the KNN algorithm. Finally, according to the comparison of average values of k minimum Euclidean distances, the classification of test samples is achieved. Experimental results are based on several real-world hyperspectral data sets, and the influence of different parameters on the classification accuracy is analyzed. Experimental results show that, with recursive filtering, the noise can be effectively removed, and the image outline can be strengthened. Compared with other hyperspectral image classification methods, the proposed method is outstanding in classification accuracy.
Wang Y M, Lang J W, Wang J Y . Status and prospect of space-borne hyperspectral imaging technology[J]. Laser and Optoelectronics Progress, 2013,50(1):75-82.
[2]
Manolakis D, Shaw G . Detection algorithms for hyperspectral imaging applications[J]. IEEE Signal Processing Magazine, 2002,19(1):29-43.
doi: 10.1109/79.974724
[3]
Bioucas-Dias J M, Plaza A, Camps-Valls G , et al. Hyperspectral remote sensing data analysis and future challenges[J]. IEEE Geoscience and Remote Sensing Magazine, 2013,1(2):6-36.
doi: 10.1109/MGRS.2013.2244672
Li Q T, Zhang L P, Yang F J , et al. The MLC’s problem in classification of hyperspectral RS image and its solving method[J]. Journal of Shandong University of Science (Natural Science Edition), 2005,24(3):61-64.
[5]
Melgani F, Bruzzone L . Classification of hyperspectral remote sensing images with support vector machines[J]. IEEE Transactions on Geoscience and Remote Sensing, 2004,42(8):1778-1790.
doi: 10.1109/TGRS.2004.831865
[6]
Fauvel M, Tarabalka Y, Benediktsson J A , et al. Advances in spectral-spatial classification of hyperspectral images[J]. Proceedings of the IEEE Transactions on Geoscience and Remote Sensing, 2013,101(3):652-675.
doi: 10.1109/JPROC.2012.2197589
Peng H T, Ke C Q . Study on object-oriented remote sensing image classification based on multi-levels segmentation[J]. Remote Sensing Technology and Application, 2010,25(1):149-154.
[9]
Vincent L, Soille P . Watersheds in digital spaces:An efficient algorithm based on immersion simulations[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1991,13(6):583-598.
doi: 10.1109/34.87344
Wang Y J, Wang Z Y, Teng Q Z , et al. Image segmentation of cutting grains based on entropy rate superpixel and region merging[J].Computer Engineering and Design, 2014(12):4223-4227.
Li X C, Zhu S A . A survey of the Markov random field method for image segmentation[J]. Journal of Image and Graphics, 2007,12(5):789-798.
[12]
Tarabalka Y Chanussot J, Benediktsson J A . Segmentation and classification of hyperspectral images using minimum spanning forest grown from automatically selected markers[J]. IEEE Transactions on Systems,Man,and Cybernetics.Part B,Cybernetics:A Publication of the IEEE Systems,Man,and Cybernetics Society, 2010,40(5):1267-1279.
doi: 10.1109/TSMCB.2009.2037132
pmid: 20051346
Cheng Z H, Xie F D . Semi-supervised classification for hyperspectral image based on spatial features and texture information[J].Bulletin of Surveying and Mapping, 2016(12):56-59,73.
Zhang Z Y, Feng W F, Zhang J H , et al. The waters feature extraction from the RS color image based on the morphological chromatic aberration[J]. Hydrographic Surveying and Charting, 2006,26(5):58-60.
[17]
Camps-Valls G, Gomez-Chova L, Munoz-Mari J , et al. Composite kernels for hyperspectral image classification[J]. IEEE Geoscience and Remote Sensing Letters, 2006,3(1):93-97.
doi: 10.1109/LGRS.2005.857031
[18]
Zhang L F, Zhang L P, Tao D C , et al. On combining multiple features for hyperspectral remote sensing image classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2012,50(3):879-893.
doi: 10.1109/tgrs.2011.2162339
[19]
Kang X D, Li S T, Benediktsson J A . Feature extraction of hyperspectral images with image fusion and recursive filtering[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014,52(6):3742-3752.
doi: 10.1109/TGRS.2013.2275613
[20]
Wang Z W, Yang J C, Nasrabadi N, et al. A max-margin perspective on sparse representation-based classification [C]//IEEE International Conference on Computer Vision.Sydney:IEEE, 2013: 1217-1224.
[21]
Chen Y, Nasrabadi N M, Tran T D . Hyperspectral image classification using dictionary-based sparse representation[J]. IEEE Transactions on Geoscience and Remote Sensing, 2011,49(10):3973-3985.
doi: 10.1109/TGRS.2011.2129595
[22]
Benediktsson J A, Palmason J A, Sveinsson J R . Classification of hyperspectral data from urban areas based on extended morphological profiles[J]. IEEE Transactions on Geoscience and Remote Sensing, 2005,43(3):480-491.
doi: 10.1109/TGRS.2004.842478
[23]
Kang X D, Li S T, Benediktsson J A . Spectral-spatial hyperspectral image classification with edge-preserving filtering[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014,52(5):2666-2677.
doi: 10.1109/TGRS.2013.2264508
[24]
Li J ,Bioucas-Dias J M,Plaza A.Hyperspectral image segmentation using a new Bayesian approach with active learning[J]. IEEE Transactions on Geoscience and Remote Sensing, 2011,49(10):3947-3960.
doi: 10.1109/TGRS.2011.2128330
2019, Vol. 31 
