Spectral-Similarity-Based Kernel of SVM for Hyperspectral Image Classification
Spectral similarity measures can be regarded as potential metrics for kernel functions, and can be used to generate spectral-similarity-based kernels. However, spectral-similarity-based kernels have not received significant attention from researchers. In this paper, we propose two novel spectral-sim...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-07-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/12/13/2154 |
| _version_ | 1797563244236242944 |
|---|---|
| author | Ke Wang Ligang Cheng Bin Yong |
| author_facet | Ke Wang Ligang Cheng Bin Yong |
| author_sort | Ke Wang |
| collection | DOAJ |
| description | Spectral similarity measures can be regarded as potential metrics for kernel functions, and can be used to generate spectral-similarity-based kernels. However, spectral-similarity-based kernels have not received significant attention from researchers. In this paper, we propose two novel spectral-similarity-based kernels based on spectral angle mapper (SAM) and spectral information divergence (SID) combined with the radial basis function (RBF) kernel: Power spectral angle mapper RBF (Power-SAM-RBF) and normalized spectral information divergence-based RBF (Normalized-SID-RBF) kernels. First, we prove these spectral-similarity-based kernels to be Mercer’s kernels. Second, we analyze their efficiency in terms of local and global kernels. Finally, we consider three hyperspectral datasets to analyze the effectiveness of the proposed spectral-similarity-based kernels. Experimental results demonstrate that the Power-SAM-RBF and SAM-RBF kernels can obtain an impressive performance, particularly the Power-SAM-RBF kernel. For example, when the ratio of the training set is <inline-formula> <math display="inline"> <semantics> <mrow> <mn>20</mn> <mo>%</mo> </mrow> </semantics> </math> </inline-formula>, the kappa coefficient of Power-SAM-RBF kernel (0.8561) is <inline-formula> <math display="inline"> <semantics> <mrow> <mn>1.61</mn> <mo>%</mo> </mrow> </semantics> </math> </inline-formula>, <inline-formula> <math display="inline"> <semantics> <mrow> <mn>1.32</mn> <mo>%</mo> </mrow> </semantics> </math> </inline-formula>, and <inline-formula> <math display="inline"> <semantics> <mrow> <mn>1.23</mn> <mo>%</mo> </mrow> </semantics> </math> </inline-formula> higher than that of the RBF kernel on the Indian Pines, University of Pavia, and Salinas Valley datasets, respectively. We present three conclusions. First, the superiority of the Power-SAM-RBF kernel compared to other kernels is evident. Second, the Power-SAM-RBF kernel can provide an outstanding performance when the similarity between spectral signatures in the same hyperspectral dataset is either extremely high or extremely low. Third, the Power-SAM-RBF kernel provides even greater benefits compared to other commonly used kernels when the sizes of the training sets increase. In future work, multiple kernels combining with the spectral-similarity-based kernel are expected to be provide better hyperspectral classification. |
| first_indexed | 2024-03-10T18:39:45Z |
| format | Article |
| id | doaj.art-a56f20f7e84e4873a50a7abe7c3aaf38 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T18:39:45Z |
| publishDate | 2020-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-a56f20f7e84e4873a50a7abe7c3aaf382023-11-20T05:57:04ZengMDPI AGRemote Sensing2072-42922020-07-011213215410.3390/rs12132154Spectral-Similarity-Based Kernel of SVM for Hyperspectral Image ClassificationKe Wang0Ligang Cheng1Bin Yong2State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing 210098, ChinaDepartment of Geographical Information Science, Hohai University, Nanjing 210098, ChinaState Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing 210098, ChinaSpectral similarity measures can be regarded as potential metrics for kernel functions, and can be used to generate spectral-similarity-based kernels. However, spectral-similarity-based kernels have not received significant attention from researchers. In this paper, we propose two novel spectral-similarity-based kernels based on spectral angle mapper (SAM) and spectral information divergence (SID) combined with the radial basis function (RBF) kernel: Power spectral angle mapper RBF (Power-SAM-RBF) and normalized spectral information divergence-based RBF (Normalized-SID-RBF) kernels. First, we prove these spectral-similarity-based kernels to be Mercer’s kernels. Second, we analyze their efficiency in terms of local and global kernels. Finally, we consider three hyperspectral datasets to analyze the effectiveness of the proposed spectral-similarity-based kernels. Experimental results demonstrate that the Power-SAM-RBF and SAM-RBF kernels can obtain an impressive performance, particularly the Power-SAM-RBF kernel. For example, when the ratio of the training set is <inline-formula> <math display="inline"> <semantics> <mrow> <mn>20</mn> <mo>%</mo> </mrow> </semantics> </math> </inline-formula>, the kappa coefficient of Power-SAM-RBF kernel (0.8561) is <inline-formula> <math display="inline"> <semantics> <mrow> <mn>1.61</mn> <mo>%</mo> </mrow> </semantics> </math> </inline-formula>, <inline-formula> <math display="inline"> <semantics> <mrow> <mn>1.32</mn> <mo>%</mo> </mrow> </semantics> </math> </inline-formula>, and <inline-formula> <math display="inline"> <semantics> <mrow> <mn>1.23</mn> <mo>%</mo> </mrow> </semantics> </math> </inline-formula> higher than that of the RBF kernel on the Indian Pines, University of Pavia, and Salinas Valley datasets, respectively. We present three conclusions. First, the superiority of the Power-SAM-RBF kernel compared to other kernels is evident. Second, the Power-SAM-RBF kernel can provide an outstanding performance when the similarity between spectral signatures in the same hyperspectral dataset is either extremely high or extremely low. Third, the Power-SAM-RBF kernel provides even greater benefits compared to other commonly used kernels when the sizes of the training sets increase. In future work, multiple kernels combining with the spectral-similarity-based kernel are expected to be provide better hyperspectral classification.https://www.mdpi.com/2072-4292/12/13/2154spectral similaritykernel functionsupport vector machinehyperspectral image |
| spellingShingle | Ke Wang Ligang Cheng Bin Yong Spectral-Similarity-Based Kernel of SVM for Hyperspectral Image Classification Remote Sensing spectral similarity kernel function support vector machine hyperspectral image |
| title | Spectral-Similarity-Based Kernel of SVM for Hyperspectral Image Classification |
| title_full | Spectral-Similarity-Based Kernel of SVM for Hyperspectral Image Classification |
| title_fullStr | Spectral-Similarity-Based Kernel of SVM for Hyperspectral Image Classification |
| title_full_unstemmed | Spectral-Similarity-Based Kernel of SVM for Hyperspectral Image Classification |
| title_short | Spectral-Similarity-Based Kernel of SVM for Hyperspectral Image Classification |
| title_sort | spectral similarity based kernel of svm for hyperspectral image classification |
| topic | spectral similarity kernel function support vector machine hyperspectral image |
| url | https://www.mdpi.com/2072-4292/12/13/2154 |
| work_keys_str_mv | AT kewang spectralsimilaritybasedkernelofsvmforhyperspectralimageclassification AT ligangcheng spectralsimilaritybasedkernelofsvmforhyperspectralimageclassification AT binyong spectralsimilaritybasedkernelofsvmforhyperspectralimageclassification |