基于Fréchet距离的光谱曲线匹配应用分析

Spectral Curve Matching Application Analysis Based on Fréchet Distance

提出了一种基于Fréchet距离的度量光谱曲线相似度的新方法,并将该方法应用于对不同植物种类的识别,对5种不同植物实测光谱曲线进行测试,并对测试结果进行了详细的分析和探讨,同时也分析了该方法在利用光谱曲线进行物种识别和匹配方面的适用性和局限性。结果表明,基于Fréchet距离的光谱曲线相似性度量精度较高。

The spectral curves of ground objects are used for the research of quantitative remote sens- ing and image classification, so it is meaningful to analyze the similarities and differences of spectral curves for each species of plant. This article puts forth a new method based on Fréchet distance to measure similarity of spectral curves used for recognition of different plant types. We tested spectral curve measured from five kinds of different plants, and present a detailed analysis and discussion of the results. We also analyzed the applicability and limitations of this method in spectral curve recogni- tion and matching application. Results showthat the precision of spectral curve similarity measurement based on Fréchet distance is much higher than other methods.