Remote sensing image fusion based on edge-preserving filtering and structure tensor
-
摘要:
高光谱（HS）遥感图像含有丰富的光谱信息，但是空间分辨率较低，而全色（PAN）遥感图像空间分辨率较高。针对高光谱遥感图像与全色遥感图像的融合问题，提出了一种新的基于边缘保持滤波和结构张量的遥感图像融合算法。首先，为了提取高光谱遥感图像的空间信息，提出使用边缘保持滤波方法，该提取方法可以保证提取的信息全部为空间细节信息，避免低频混叠。其次，对全色遥感图像采用高斯-拉普拉斯图像增强算法进行图像锐化，降低图像噪声，锐化细节信息。再次，为得到总空间信息，提出使用结构张量的自适应加权策略。传统的融合算法通常仅从全色遥感图像中提取空间信息，可能会引起光谱失真或空间细节加入不足等问题，为了克服这些问题，提出的自适应加权策略得到的总空间信息不仅包含全色遥感图像的空间信息，还包含高光谱遥感图像的空间信息，且自适应加权相对于全局常数加权，可以自动选取更加合适的加权数据。最后，通过构建可以控制光谱和空间失真的增益矩阵，将总空间信息注入到插值的高光谱遥感图像的每个波段中，得到融合的高光谱图像。实验结果表明，本文提出的遥感图像融合算法，在客观评价方面，取得了最优的空间和光谱性能，在视觉效果上，与其他融合算法相比，可以更有效地提高空间分辨率和保持光谱信息。
Abstract:The hyperspectral (HS) remote sensing image which contains abundant spectral information generally has low spatial resolution. While the panchromatic (PAN) remote sensing image has high spatial resolution. In order to fuse the HS and PAN remote sensing images, a new fusion algorithm based on edge-preserving filtering and structure tensor is proposed. First, to avoid low-frequency aliasing, an edge-preserving filter is introduced to extract the spatial information of the HS image. In order to sharpen the spatial information of the PAN image, an image enhancement approach is applied to the PAN image. Then, an adaptive weighting strategy which is based on the structure tensor is proposed to obtain the total spatial information. The presented adaptive weighting strategy which is different from the traditional fusion method reduces the spectral distortion and provides adequate spatial information. The injection matrix is finally constructed to reduce spectral and spatial distortion, and the fused image is generated by injecting the complete spatial information. Experimental results demonstrate that the proposed method provides more spatial information and preserves more spectral information compared with the state-of-art fusion methods.
-
图 1 高光谱图像融合算法框图
Figure 1. Block diagram of proposed hyperspectral image fusion algorithm
图 2 不同融合算法得到的Pavia University数据集的融合结果
Figure 2. Fusion results of different fusion algorithms for Pavia University dataset
图 3 不同融合算法得到的Moffett field数据集的融合结果
Figure 3. Fusion results of different fusion algorithms for Moffett field dataset
图 4 不同融合算法得到的Washington DC数据集的融合结果
Figure 4. Fusion results of different fusion algorithms for Washington DC dataset
表 1 边缘保持滤波方法对各数据集的客观评价指标
Table 1. Objective evaluation indices of edge-preserving filtering method for each dataset
数据集 算法 CC SAM RMSE ERGAS Pavia University GSI 0.930 4 6.774 0 0.044 9 4.439 9 本文 0.930 4 6.568 6 0.038 5 4.148 6 Moffett field GSI 0.954 3 6.355 0 0.030 2 4.052 2 本文 0.965 0 6.347 0 0.030 1 4.030 9 Washington DC GSI 0.866 2 7.233 0 0.013 1 74.957 2 本文 0.879 4 7.232 0 0.013 3 73.674 9 
下载: 导出CSV
表 2 基于结构张量的自适应加权策略对各数据集的客观评价指标
Table 2. Objective evaluation indices of structure tensor based adaptive weighting strategy for each dataset
数据集 算法 CC SAM RMSE ERGAS Pavia University GFP 0.932 6 6.593 2 0.042 5 4.235 8 GFGW 0.906 6 6.766 4 0.041 2 4.783 2 GFPL 0.925 6 6.586 4 0.039 4 4.257 1 本文 0.930 4 6.568 6 0.038 5 4.148 6 Moffett field GFP 0.957 5 6.347 0 0.031 4 4.351 8 GFGW 0.956 6 6.358 0 0.031 0 4.337 2 GFPL 0.962 3 6.348 0 0.030 5 4.031 5 本文 0.965 0 6.347 0 0.030 1 4.030 9 Washington DC GFP 0.873 8 7.232 1 0.013 6 73.695 0 GFGW 0.865 9 7.232 5 0.014 0 76.223 0 GFPL 0.874 5 7.232 4 0.013 4 74.326 6 本文 0.879 4 7.232 0 0.013 3 73.674 9
下载: 导出CSV
表 3 Pavia University数据集的融合结果客观评价指标
Table 3. Objective evaluation indices of fusion results for Pavia University dataset
算法 CC SAM RMSE ERGAS PCA 0.923 4 7.656 6 0.040 8 4.783 0 GFPCA 0.795 2 9.449 5 0.061 6 7.135 8 GSA 0.921 9 8.799 2 0.038 9 4.532 1 BSR 0.900 3 8.676 6 0.044 1 5.301 6 MGH 0.930 2 6.909 0 0.038 9 4.235 5 本文 0.930 4 6.568 6 0.038 5 4.148 6
下载: 导出CSV
表 4 Moffett field数据集的融合结果客观评价指标
Table 4. Objective evaluation indices of fusion results for Moffett field dataset
算法 CC SAM RMSE ERGAS PCA 0.905 0 12.425 5 0.047 5 6.698 0 GFPCA 0.915 7 10.319 8 0.044 1 6.287 6 GSA 0.949 7 8.660 5 0.036 1 5.044 4 BSR 0.954 0 8.037 2 0.032 3 4.713 1 MGH 0.964 4 6.007 8 0.032 5 4.356 8 本文 0.965 0 6.347 0 0.030 1 4.030 9
下载: 导出CSV
表 5 Washington DC数据集的融合结果客观评价指标
Table 5. Objective evaluation indices of fusion results for Washington DC dataset
算法 CC SAM RMSE ERGAS PCA 0.853 2 7.361 9 0.013 6 83.414 5 GFPCA 0.768 9 9.932 2 0.013 9 67.944 1 GSA 0.870 1 7.258 0 0.018 4 83.997 9 BSR 0.826 9 10.012 5 0.013 8 77.749 1 MGH 0.877 7 7.261 8 0.015 6 79.859 8 本文 0.879 4 7.232 0 0.013 3 73.674 9
下载: 导出CSV
-
[1] LI Y S, HU J, ZHAO X, et al. Hyperspectral image super-resolution using deep convolutional neural network[J]. Neurocomputing, 2017, 266:29-41. doi: 10.1016/j.neucom.2017.05.024 [2] MOOKAMBIGA A, GOMATHI V. Comprehensive review on fusion techniques for spatial information enhancement in hyperspectral imagery[J].Multidimensional Systems and Signal Processing, 2016, 27(4):863-889. doi: 10.1007/s11045-016-0415-2 [3] TU T M, SU S C, SHYU H C, et al. A new look at IHS-like image fusion methods[J].Information Fusion, 2001, 2(3):177-186. doi: 10.1016/S1566-2535(01)00036-7 [4] CHAVEZ P S, KWARTENG A Y.Extracting spectral contrast in Landsat thematic mapper image data using selective principal component analysis[J].Photogrammetric Engineering and Remote Sensing, 1989, 55(3):339-348. [5] LABEN C, BROWER B.Process for enhancing the spatial resolution of multispectral imagery using pan-sharpening: United States Patent 6011875[P].2000-01-04. [6] QU J H, LI Y S, DONG W Q.Hyperspectral pansharpening with guided filter[J].IEEE Geoscience and Remote Sensing Letters, 2017, 14(11):2152-2156. doi: 10.1109/LGRS.2017.2755679 [7] MALLAT S.A theory for multiresolution signal decomposition:The wavelet representation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 1989, 11(7):674-693. doi: 10.1109/34.192463 [8] VIVONE G, RESTAINO R, MAURO D M, et al.Contrast and error-based fusion schemes for multispectral image pansharpening[J].IEEE Geoscience and Remote Sensing Letters, 2014, 11(5):930-934. doi: 10.1109/LGRS.2013.2281996 [9] LIU J G.Smoothing filter based intensity modulation:A spectral preserve image fusion technique for improving spatial details[J].International Journal of Remote Sensing, 2000, 21(18):3461-3472. doi: 10.1080/014311600750037499 [10] YOKOYA N, YAIRI T, IWASAKI A.Coupled nonnegative matrix factorization unmixing for hyperspectral and multispectral data fusion[J].IEEE Transactions on Geoscience and Remote Sensing, 2012, 50(2):528-537. doi: 10.1109/TGRS.2011.2161320 [11] SIMOES M, DIAS J B, ALMEIDA L B, et al.A convex formulation for hyperspectral image superresolution via subspace-based regularization[J].IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(6):3373-3388. doi: 10.1109/TGRS.2014.2375320 [12] WEI Q, DIAS J M, DOBIGEON N, et al.Hyperspectral and multispectral image fusion based on a sparse representation[J].IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(7):3658-3668. doi: 10.1109/TGRS.2014.2381272 [13] HE K, SUN J, TANG X.Guided image filtering[J].IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(6):1397-1409. doi: 10.1109/TPAMI.2012.213 [14] HARRIS C.A combined corner and edge detector[C]//Proceedings of the Alvey Vision Conference, 1988: 147-151. [15] WALD L, RANCHIN T, MANGOLINI M.Fusion of satellite images of different spatial resolutions:Assessing the quality of resulting images[J].Photogrammetric Engineering and Remote Sensing, 1997, 63(6):691-699. [16] LONCAN L, ALMEIDA L B, DIAS J M, et al.Hyperspectral pansharpening:A review[J].IEEE Geoscience Remote Sensing Magazine, 2015, 3(3):27-46. doi: 10.1109/MGRS.2015.2440094 [17] ZHANG L, ZHANG L, TAO D, 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 [18] AIAZZI B, BARONTI S, SELVA M.Improving component substitution pansharpening through multivariate regression of MS+pan data[J].IEEE Transactions on Geoscience and Remote Sensing, 2007, 45(10):3230-3239. doi: 10.1109/TGRS.2007.901007 [19] LIAO W, HUANG X, COILLIE F, et al.Processing of multiresolution thermal hyperspectral and digital color data:Outcome of the 2014 IEEE GRSS data fusion contest[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2015, 8(6):2984-2996. doi: 10.1109/JSTARS.2015.2420582 -
下载:
点击查看大图
计量
- 文章访问数: 1107
- HTML全文浏览量: 143
- PDF下载量: 389
- 被引次数: 0
