|
|||||
|
|
| 结合KPCA和分形维提取高光谱遥感影像特征的方法 | |
| 引用本文： | 沈照庆,黄亮,陶建斌.结合KPCA和分形维提取高光谱遥感影像特征的方法[J].测绘科学,2012,37(5):27-29,42. |
| 作者姓名： | 沈照庆 黄亮 陶建斌 |
| 作者单位： | 1. 长安大学公路学院,西安,710064 2. 中煤航测遥感局,西安,710054 3. 武汉大学遥感信息工程学院,武汉,430079 |
| 基金项目： | 国家重点基础研究发展计划(973)项目“对地观测数据-空间信息-地学知识的转化机理”资助项目(2006CB701303);中央高校基本科研业务专项资金项目“基于SVM高光谱影像道路提取与分析研究”(CHD2011JC011) |
| 摘 要： | 本文将KPCA和分形维有机结合,进行高光谱影像特征提取,实现优势互补:选择合适的核函数和分形维计算方法,设计了3种组合算法,优化了特征提取效果,并对AVIRIS实验结果进行了分析评价,结果显示在相同条件下,SVM的分类精度要高于其他分类算法,KPCA+Fractal特征提取更有利于地物的分类识别。 |
| 关 键 词： | 高光谱遥感影像 核函数 核PCA 分形维 特征提取 |
Extraction of hyperspectral RS image feature with KPCA and fractal dimension |
|
| SHEN Zhao-qing , HUANG Liang , TAO Jian-bin.Extraction of hyperspectral RS image feature with KPCA and fractal dimension[J].Science of Surveying and Mapping,2012,37(5):27-29,42. | |
| Authors: | SHEN Zhao-qing HUANG Liang TAO Jian-bin |
| Institution: | ③(①School of Highway,Chang’an University,Xi’an 710064,China;②Aerial Photogrammetry and Remote Sensing Coal Reconnaissance Institute,Xi’an 710054,China;③School of Remote Sensing and Information Engineering,Wuhan University,Wuhan 430079,China) |
| Abstract: | In this paper,the hyperspectral RS image feature was extracted with the combination of fractal dimension and KPCA,both of them got complementary advantages.Three optimization algorithms were designed,in which the appropriate kernel function and fractal dimension calculation methods were designed.It optimized the feature extraction,and the AVIRIS results were evaluated at last.The results showed that under the same conditions,SVM classification accuracy is higher than other classification algorithm,the feature extraction using KPCA+Fractal could be more conducive to surface features of the classification. |
| Keywords: | hyperspectral RS image kernel function Kernel PCA(KPCA) fractal dimension feature extraction |
| 本文献已被 CNKI 万方数据 等数据库收录！ | |