基于多时相和多角度光谱信息的作物株型遥感识别初探

作物群体具有一定的冠层几何结构(株型)，对于不同株型的品种，在相同的叶面积指数时冠层反射光谱往往不同，使得利用冠层反射光谱来反演叶面积指数等生物物理和生物化学参数时存在不同株型产生的误差，该文定量研究了不同叶面积指数条件下，作物株型对冠层反射光谱的影响，并提出运用波长800 nm处起身期的冠层反射光谱与该波长处拔节期和起身期冠层反射光谱的比值，可以初步实现高密度披散型品种、低密度披散型品种、高密度中间型品种、低密度中间型品种、高密度直立型品种和低密度直立型品种的遥感识别，结合一定条件下选取的15°、30°和45°观测天顶角下，与可见光和近红外波段(波长)处的二向反射冠层反射光谱数值大小进行结合，可以初步实现作物株型的遥感识别。

Remote sensing identification of plant structural types based on multi-temporal and bidirectional canopy spectrum

Investigations have been made on identification of erective, middle and loose varieties by multi-temporal and bidirectional canopy reflected spectrum. It indicates that canopy reflected spectrum has significant differences between reviving stage and jointing stage. Canopy reflectance at 800 nm was similar among erective, middle and loose varieties at erecting stage, which has significant differences for the ratio of canopy reflectance at 800 nm at jointing stage to the same wavelength of canopy reference at erecting stage (Ratio). The value of ratio was: loose varieties>middle varieties> erective varieties. The combination of ratio and canopy reflectance at 800 nm at erecting stage could be used for the identification of plant structural types based on the multi-temporal data. The reflectance of visible and near infrared bands at 15°, 30°and 45° field of view for the main viewing plane could be used for the identification of plant structural types based on bidirectional data. So, it is feasible to identify loose, middle and erective varieties of wheat by multi-temporal and bidirectional canopy reflected spectrum.