Retrieval of chlorophyll content and LAI of crops using hyperspectral techniques: application to PROBA/CHRIS data
Retrieval of chlorophyll content and LAI of crops using hyperspectral techniques: application to PROBA/CHRIS data
复制标题
DOI：
10.1080/01431160802238401
复制
发表时间：
2008-01-01
影响因子：
3.4
通讯作者：
Moreno, J.
中科院分区：
文献类型：
--
作者：
Delegido, J.;Fernandez, G.;Moreno, J.
高光谱/多角度数据可以在冠层水平检索重要的植被特性，如叶面积指数（LAI）和叶片叶绿素含量。目前的方法是基于生物物理性质与在考虑的光谱范围内存在特定吸收特征的光谱波段（从完整的高光谱/多角度信息）检索之间的关系。此外，新的传感器如PROBA/CHRIS提供连续的高光谱反射率测量，可以被认为是波长的连续函数。这些连续函数的数学分析为利用更强大、更稳定的数学工具来研究光谱反射率与生物物理变量之间的关系提供了一种新的方法，特别是在反演LAI和叶绿素含量方面。在欧洲航天局（ESA）光谱巴拉克斯运动（SPARC）实验的总体背景下，在西班牙拉曼查进行了广泛的实地研究，同时对机载成像光谱仪（AHS, HyMAP， ROSIS）和克里斯- proba和MERIS传感器进行了飞越。在SPARC-2003和SPARC-2004运动期间，在Barrax研究区进行了大量地面测量（包括LAI， fCover，叶片叶绿素a+b，叶片含水量和叶片生物量）以及其他补充数据，共获得了17张chriss - proba图像。从总共九种不同的作物中选择了具有代表性的点，并从田间活动几天内获得的克里斯- proba图像中检索。分析了来自5个不同观测角度的约250个反射光谱。在500nm和750nm之间，利用三次多项式函数对高光谱反射光谱进行了调整，观察到LAI值与这些多项式系数之间的相关性，通过数学拟合得到LAI。另一方面，光谱反射曲线下的面积在600nm至700nm的区间内计算，在红光光谱区间内，叶绿素的吸收特征很强，曲线下的面积也与作物的叶绿素含量有很强的相关性。此外，这些区域与叶绿素含量之间存在线性关系。这种关系使得基于光谱信息的卫星数据检索叶片叶绿素成为可能。这两种方法几乎与所采用的观测角度无关。对卫星立交桥同时获得的大量现场测量数据和广泛的可用数据范围，使这两种方法能够以大量数据和统计一致的方式得到验证。
Hyperspectral/multiangular data allow the retrieval of important vegetation properties at canopy level, such as the Leaf Area Index (LAI) and Leaf Chlorophyll Content. Current methods are based on the relationship between biophysical properties and retrievals from those spectral bands (from the complete hyperspectral/multiangular information) where specific absorption features are present within the considered spectral range. Furthermore, new sensors such as PROBA/CHRIS provide continuous hyperspectral reflectance measurements that can be considered as a continuous function of wavelength. The mathematical analysis of these continuous functions allows a new way of exploiting the relationships between spectral reflectance and biophysical variables by more powerful and stable mathematical tools, in particular for the retrieval of LAI and chlorophyll content. Within the overall context of the European Space Agency (ESA) Spectra Barrax Campaign (SPARC) experiment, an extensive field study was carried out in La Mancha, Spain, simultaneously to the overflight of airborne imaging spectrometers (AHS, HyMAP, ROSIS) and the overpass of CHRIS-PROBA and MERIS sensors. During the SPARC-2003 and SPARC-2004 campaigns, numerous ground measurements were made in the Barrax study area (covering LAI, fCover, leaf chlorophyll a+b, leaf water content and leaf biomass), together with other complementary data, and a total of 17 CHRIS-PROBA images were acquired. Representative points have been selected from a total of nine different crops, and also retrieved from the CHRIS-PROBA images acquired within the days of the field campaign. About 250 reflectance spectra from five different observation angles have been analysed. Hyperspectral reflectance spectra have been adjusted by means of third-degree polynomial functions between 500nm and 750nm, and correlations observed between LAI values and the coefficients of these polynomials yielded LAI as a result of the mathematical fitting. On the other hand, the area under the spectral reflectance curves has been calculated in the interval from 600nm to 700nm, the region of the red spectral interval where strong absorption features for chlorophyll have been observed, though areas under the curves are also strongly correlated to the chlorophyll content of the crops. Furthermore, a linear relationship between these areas and the chlorophyll content is proposed in this work. This relationship allows the retrieval of leaf chlorophyll by satellite data, based on the spectral information. Both of the proposed methods are almost independent of the observation angles employed. The high number of in situ measurements acquired simultaneously to satellite overpasses, and the broad available range of data, have allowed validation of both methods, with a large number of data and in a statistically consistent manner.