基于最佳波段指数和J-M距离可分性的高光谱数据最佳波段组合选取研究——以环境小卫星高光谱数据在东莞市的应用为例

Determination on the Optimum Band Combination of HJ-1A Hyperspectral Data in the Case Region of Dongguan Based on Optimum Index Factor and J-M Distance

高光谱数据普遍存在波段相关性强、数据冗余严重的特点,因此选择合适的波段组合,是高效开展后续应用研究的基础。以东莞市为研究区,应用环境与灾害监测预报小卫星(HJ-1A)高光谱数据,在分析各波段信息含量和波段间相关性的基础上,使用了3种经典的波段指数选择最佳波段组合;针对经典模型应用中存在的问题,对最佳指数模型进行了改进,通过对波段均方差和相关性设置一定阈值,筛选得到一个较为合理的波段组合;最后,针对草地、林地和耕地3种地物,应用J-M距离模型对3种地物的可分性进行判别,并指出:50-80-108波段组合,50-79-108波段组合以及50-80-111波段组合是分别用于草地—林地、草地—耕地、耕地—林地分类的最佳波段组合。

Comparing with traditional remote sensing data,strong band relevant and data redundancy are prevalent in the hyperspectral remote sensing data,so band selection is necessary for the efficient research.This paper taking Dongguan as a study area,using Environment and Disaster Monitoring and Forecasting of small satellite hyperspectral data,comprehensively analyzed the information content and correlation of the bands,then using three kinds of classical band index to choose the best band combinations.Through contrastive analysis of the three models,we improved the Optimum Index Factor by giving thresholds to coefficient of correlation and mean square deviation to select the optimum band combinations on the problems in the classical models.In the end,taking the farmland,woodland and grassland for example,apply Jeffreys-Matusita Distance model to calculate their separability and pointed out that the band combination of 50-80-108、50-79-108 and 50-80-111 are the best combinations for distinguishing grassland-woodland,grassland-farmland and woodland-farmland separately.