盐生植物作为干旱区绿洲重要的生命支持系统, 对于维持绿洲生态系统平衡起着核心的作用。很多学者对此展开了大量研究, 但是对盐生植物的光谱特征研究较少, 因此, 该文以渭干河-库车河三角洲绿洲盐漠带典型盐生植物为研究对象, 利用FieldSpec Pro FR便携式地物波谱仪, 2010年10月通过对盐穗木(Halostachys caspica)、盐节木(Halocnemum strobilaceum)、骆驼刺(Alhagi sparsifolia)、白刺(Nitrarria sibirica)、柽柳(Tamarix ramosissima)、花花柴(Karelinia caspia)及芦苇(Phragmites australis)野外光谱数据采集和处理, 获得了这7种盐生植物的高光谱数据, 然后采用光谱学分析方法分析其特征变化, 结果表明: (1)采用Percentile Filter平滑方法进行光谱噪声去除, 有很好的去除噪音效果。(2)对3种类型盐生植物样本的原始光谱曲线提取光谱吸收特征参数, 结果发现稀盐盐生植物、泌盐盐生植物、拒盐盐生植物的波谷波长位置接近, 说明盐生植物的吸收波段特征有一定的相似性。(3)对盐生植物光谱反射曲线进行归一化处理, 发现在消除土壤等背景影响、突出目标及消除同物异谱现象上具有很好的效果, 能够提高盐生植被的识别精度。(4)应用二阶导数的方法得到盐生植物识别的9个最佳波段: 510、550、690、730、950、1 150、1 210、1 290和1 310 nm。研究成果不仅可以为盐漠带植物的高光谱遥感数据处理提供一定的科学依据, 而且可以为盐漠带盐生植物的遥感识别和分类提供一定的参考。
Aims As one of the main components of terrestrial ecosystems, halophytes play a key role in maintaining arid zone ecosystem balance. Remote sensing has potential for monitoring halophytes in large areas. However, few studies on the spectral properties of halophytes are available. The spectral features of halophytes on the ground provide an important basis for remote sensing applications. Methods This study used the FieldSpec Pro FR portable spectroscope to record ground reflectance of halophytes, using typical halophytes in the salinization-desert zone of a delta oasis in Weigan-Kuqa River as an example. We first analyzed the data noise characteristic of the spectrum. The noise was removed using the Percentile Filter method. We employed differential spectrum technology to reduce the influence of environmental background. Second, we extracted parameters of spectral absorption features for typical halophytes: Halostachys caspica, Halocnemum strobilaceum, Alhagi sparsifolia, Nitrarria sibirica, Tamarix ramosissima, Karelinia caspia and Phragmites australis. Third, we processed spectral reflectance curves of halophytes using a normalization method. Important findings The wave troughs/wave length positions of euhalophyte, secretohalophyte and pseudohalophy are very similar, indicating that halophytes have similar absorption band characteristics. Elimination of the elements of noises and deflections caused by different scanning conditions such as soil background improved the recognition accuracy of halophytes. We ultimately identified nine optimal spectral bands, i.e., 510, 550, 690, 730, 950, 1 150, 1 210, 1 290 and 1 310 nm, that appeared to contain majority of the salinization-desert zone information of the multispectral signatures. The nine bands can be used to differentiate salinization-desert zone halophytes types. The results not only provide a scientific basis for hyperspectral remote sensing image processing, but also supply references for the identification and classification of halophytes using remote sensing technology.