内蒙古荒漠化草原植被高光谱特征提取与分析

Extraction and analysis of hyperspectral characteristics of desert vegetation of Inner Mongolia

针对内蒙古荒漠化草原植被光谱缺少定量参数分析比较,提出了将微分法和包络线去除法相结合的方法,对植被具有诊断性的红边(680～750nm)和差异性较大的近红外波段(550～760nm)进行特征提取,并进行定量对比分析。利用GaiaSky-mini型机载地面两用型高光谱仪对荒漠化草原的建群种短花针茅、优势种冷蒿、退化指示种猪毛菜进行实地高光谱数据采集,并对数据进行降噪平滑、一阶微分、二阶微分和包络线去除法等处理,还对3种不同植被的光谱特征、红边参数特征和吸收峰面积等进行了分析。研究表明,短花针茅的反射率最低,冷蒿的红边效应最为明显,吸收峰面积及右面积差异可以将猪毛菜与冷蒿识别出来。所得结论可为荒漠植被识别及无人机遥感反演提供参考。

As the lack of quantitative analysis and comparison on vegetation spectrum of Inner Mongolia desert grassland,this paper proposes a method which combines differential method and envelope removal method to extract characteristics from diagnostic red edge band of vegetation (680~750 nm) and large difference near infrared band (550~760 nm).Then the parameters were quantitatively and comparatively analyzed.In the research area of desert steppe of Siziwang Banner, the hyperspetral data of stipa breviflora(establishment species),artemisia frigida(dominant species),and salsola(degraded indicator species) were collected by the GaiaSky-mini airborne and terrestrial hyperspectral instrument.Then the data were processed by noise reduction and smoothing,first-order differentiation,second-order differentiation and envelope removal.Moreover,the processed results were used for comparative analysis,red edge parameter characteristics and absorption peak area analysis.Research results suggest that stipa breviflora is the lowest of the three,the red edge effect of artemisia frigida is the most obvious,and the difference between the absorption peak area and the right area can be identified in salsola and artemisia frigida.The conclusion is of great significance to the recognition of desert vegetation and remote sensing inversion of unmanned aerial vehicle.