典型地物实测光谱的相似性测度与实验分析

Spectral Similarity Measure and Experimental Analyses for Field Spectroscopy

摘要:

对于高光谱向量非线性空间,提出了基于Hausdorff距离的地物光谱相似性测度模型,并对模型的适用性进行讨论.此外通过对试验区域由高分辨率光谱仪实地测得的地物光谱数据进行实验分析,得到各种相似性测度模型的匹配聚类结果,并对其有效性和精度进行评估.

Abstract:

A new spectral similarity measure based on Hausdorff distance, a nonlinear operator, which measures the mismatch of the two sets, was proposed and experimented.The performance of different spectral similarity measures was compared by using real field spectrometric data. Statistical methods were also used to assess the measurement accuracy and validity.The experiment result shows that the similarity measure based on Hausdorff distance is more effective than others in spectral matching aspect.