摘要
提出了一种新型光谱相似性测度及其参数的自适应选择方法,并且将其应用到了高光谱影像地物检测中。由于这种相似性测度基于光谱角度余弦(SAC),因此在理论上对因光照强度变化、阴影和遮挡等引起的同种地物光谱变化的适应性较强。最后利用两幅高光谱影像进行了实验分析,实验结果证明提出的方法不仅能扩大阈值取值区间,而且可提高检测的精度。
A novel spectral similarity measurement and its adaptive selection of parameter were proposed and applied to object detection in hyperspectral imagery.As the novel similarity measurement is based on spectral angel cosine（SAC）,in theory,it has a good adjustability for spectral curve variation for the same materials coming from radiation intensity variations,shadow,and shading etc.Lastly,the experiments were carried on with two hyperspectral images,and the results of the experiments proved that the proposed method could not only extend the threshold area coverage but also improve the precision of detection.
出处
《测绘科学技术学报》
北大核心
2012年第1期42-46,共5页
Journal of Geomatics Science and Technology
基金
国家自然科学基金资助项目(41072248)
关键词
高光谱影像
地物检测
光谱相似性测度
参数自适应选择
光谱角度余弦
hyperspectral image
object detection
spectral similarity measurement
adaptive selection of parameter
SAC（Spectral Angel Cosine）
