| 基于波段组合的高光谱数据湿地分类研究 |
| |
| 引用本文： | 乔雯钰,龙亦凡,付杰.基于波段组合的高光谱数据湿地分类研究[J].北京测绘,2020(5):651-656. |
| |
| 作者姓名： | 乔雯钰 龙亦凡 付杰 |
| |
| 作者单位： | 山东科技大学测绘科学与工程学院 |
| |
| 摘 要： | 对高光谱数据进行波段组合,可以减少信息量的冗余,提高数据的处理速度。对黄河口入海口湿地进行分类,对合理利用、开发保护该地区湿地资源具有重要意义。本文首先分析了“珠海一号”高光谱数据各个波段的信息量及波段之间的相关系数,然后利用最佳波段指数(OIF)方法选出波段组合B7-B8-B32,进一步在OIF基础上设置信息量与相关系数阈值,选出波段组合B7-B18-B32,实验结果证明分类精度提高了5.4%。最后,根据地物的光谱特征分析,选择光谱差异较大的波段进行组合B6-B13-B18,分类后精度比OIF筛选出的波段组合精度高12.6694%。经实验验证,结合地物光谱特征的波段组合可以大大提高分类精度。
|
| 关 键 词： | 高光谱数据 最佳波段指数 湿地分类 黄河入海口湿地 地物光谱特征 |
Study on Wetland Classification of Hyperspectral Data Based on Band Combination |
| |
| QIAO Wenyu,LONG Yifan,FU Jie.Study on Wetland Classification of Hyperspectral Data Based on Band Combination[J].Beijing Surveying and Mapping,2020(5):651-656. |
| |
| Authors: | QIAO Wenyu LONG Yifan FU Jie |
| |
| Institution: | (College of Geomatics,Shandong University of Science and Technology,Qingdao Shandong 266510,China) |
| |
| Abstract: | The Yellow River estuary wetland is currently facing threats such as pollution and seawater intrusion.The wetland area is shrinking seriously and the ecological function is gradually degraded.Therefore,the classification of the Yellow River estuary wetland is of great significance for the rational use and development of wetland resources in the region.This paper first analyzes the information volume and the correlation coefficient between the bands of the“Zhuhai No.1”hyperspectral data,which provides a reference for the selection of the wetland classification band.Then use the Optimum Index Factor(OIF)method to select the band combination as B7-B8-B32 to reduce the dimensionality of the hyperspectral data.Further,based on the OIF,the information amount and the correlation coefficient threshold are set,and the band with less information and high correlation is eliminated.The selected band combination is B7-B18-B32,and the experimental result proves that the classification accuracy is improved by 5.4%.Finally,according to the spectral characteristics analysis of the features,the bands with larger spectral differences are selected to combine B6-B13-B18,and the accuracy of classification after classification is 12.6664% higher than that of OIF. |
| |
| Keywords: | hyperspectral data optimum index factor wetland classification Yellow River estuary wetland spectral characteristics of features |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
