荔枝关键生育期冠层氮素动态变化的遥感监测

Agricultural applications are far more demanding in terms of spatial resolution、timeliness and practicability compared to other remote sensing applications. In a regional scale, the low / medium resolution images and hyperspectral images are limited in practical use because of the coarse resolution and the costs, respectively. The multispectral images with red-edge band present some advantages. In our early study, we found that during different growth stages, there was significant difference of litchi canopy reflectance in response to the changes of nitrogen content. This study will focus on the dynamics monitoring of litchi canopy nitrogen in critical growth periods by remote sensing and plan fertilizer application of litchi orchard accordingly. The main research contents include the followings. Firstly, the spectral library will be built based on litchi canopy reflectance and nitrogen content in key growth stages. Secondly, the relationship between spectral data and nitrogen will be also investigated, and the canopy reflectance will be simulated by PROSAIL model. Thirdly, the relationship between simulated data and nitrogen content will be analyzed indirectly by the interrelationship between chla and nitrogen. Finally the effective and practical nitrogen retrieval model by remote sensing image will be developed based on the analysis.

农业遥感应用对遥感影像空间分辨率、时效性、可用性有很高的要求。在区域尺度氮素反演中，中低分辨率遥感影像因分辨率粗糙应用受限，高光谱影像则受限于获取成本较高和数据冗余，而具有红边波段的高空间分辨率多光谱影像正显示出成本低，可用性好的优点。我们早期的研究发现荔枝不同生育期的冠层光谱对氮素含量变化响应有显著差别。本研究拟开展荔枝年生长周期关键生育期的冠层反射光谱与氮素关系分析，在此基础上利用具有红边波段的高空间分辨率多光谱遥感影像对荔枝年生长周期的关键生育期氮素变化进行时空动态监测。主要内容包括基于荔枝年生长周期的关键生育期实测冠层光谱数据与氮素，建荔枝冠层光谱-氮素波谱库，一方面分析氮素与光谱的关系；另一方面利用波谱库和PROSAIL物理模型，模拟像素尺度冠层光谱随叶绿素变化的响应，利用叶绿素间接研究氮素变化；最终构建考虑荔枝生育期的有效的可在实践中应用的氮素动态遥感反演模型。

本课题研究了荔枝年生长周期关键生育期的冠层反射光谱与氮素关系，并分析了部分冠层结构参数和生化参数对氮素反演的影响。发现作物品种，采样时间等对氮素反演精度有显著的影响；利用包络线去除建立的差值比值指数能够很好的反演各采样期的氮素含量；各生育期氮素反演模型选用的波段并不一致，具体选用的波段与氮素和叶绿素、蛋白质、水分等的相关程度高低有关。5个采样期中，果实成熟期氮素反演精度最高，花芽分化期氮素反演精度最低。同时模拟7种多光谱传感器的波段，发现在红边范围内，波段数据较多的传感器在LAI反演中精度更高。