Retrieval of aboveground crop nitrogen content with a hybrid machine learning method

Abstract

Hyperspectral acquisitions have proven to be the most informative Earth observation data source for the estimation of nitrogen (N) content, which is the main limiting nutrient for plant growth and thus agricultural production. In the past, empirical algorithms have been widely employed to retrieve information on this biochemical plant component from canopy reflectance. However, these approaches do not seek for a cause-effect relationship based on physical laws. Moreover, most studies solely relied on the correlation of chlorophyll content with nitrogen, and thus neglected the fact that most N is bound in proteins. Our study presents a hybrid retrieval method using a physically-based approach combined with machine learning regression to estimate crop N content. Within the workflow, the leaf optical properties model PROSPECT-PRO including the newly calibrated specific absorption coefficients (SAC) of proteins, was coupled with the canopy reflectance model 4SAIL to PROSAIL-PRO. The latter was then employed to generate a training database to be used for advanced probabilistic machine learning methods: a standard homoscedastic Gaussian process (GP) and a heteroscedastic GP regression that accounts for signal-to-noise relations. Both GP models have the property of providing confidence intervals for the estimates, which sets them apart from other machine learners. Moreover, a GP-based sequential backward band removal algorithm was employed to analyze the band-specific information content of PROSAIL-PRO simulated spectra for the estimation of aboveground N. Data from multiple hyperspectral field campaigns, carried out in the framework of the future satellite mission Environmental Mapping and Analysis Program (EnMAP), were exploited for validation. In these campaigns, corn and winter wheat spectra were acquired to simulate spectral EnMAP data. Moreover, destructive N measurements from leaves, stalks and fruits were collected separately to enable plant-organ-specific validation. The results showed that both GP models can provide accurate aboveground N simulations, with slightly better results of the heteroscedastic GP in terms of model testing and against in situ N measurements from leaves plus stalks, with root mean square error (RMSE) of 2.1 g/m². However, the inclusion of fruit N content for validation deteriorated the results, which can be explained by the inability of the radiation to penetrate the thick tissues of stalks, corn cobs and wheat ears. GP-based band analysis identified optimal spectral settings with ten bands mainly situated in the shortwave infrared (SWIR) spectral region. Use of well-known protein absorption bands from the literature showed comparative results. Finally, the heteroscedastic GP model was successfully applied on airborne hyperspectral data for N mapping. We conclude that GP algorithms, and in particular the heteroscedastic GP, should be implemented for global agricultural monitoring of aboveground N from future imaging spectroscopy data.

A Survey of Active Learning for Quantifying Vegetation Traits from Terrestrial Earth Observation Data

“…Data were collected at the following dates at the wheat field: 29/3, 10/4, 10/5, 29/5, 13/6, 26/6, 6/7 and 17/7 in 2017 and 12/4, 18/4, 27/4, 7/5, 5/6, 21/6 and 13/7 in 2018. For corn, sampling was done at the following dates: 13 [10]. Further, the studies by Danner et al [102] and Wocher et al [99] inform about sampling design, size and location of ESUs, as well as measurements of other biochemical and biophysical variables.…”

“…In the last five decades, numerous retrieval methods have been proposed and developed to predict biophysical and biochemical vegetation traits from EO data, ranging from parametric and nonparametric regressions to physically-based and hybrid approaches [5][6][7]. Since these studies provide exhaustive and up-to-date taxonomies of quantitative retrieval methods, we will concentrate here on the recently promoted hybrid retrieval workflows [8][9][10][11][12]. Hybrid retrieval strategies denominate a combination of radiative transfer models (RTM), providing physical constraints and domain knowledge [13], with fast and flexible machine learning (ML) regression algorithms.…”

A Novel Machine Learning Approach to Estimate Grapevine Leaf Nitrogen Concentration Using Aerial Multispectral Imagery

“…The emergence of advanced machine learning techniques, along with high-performance computational power, has provided new opportunities to translate image-based datasets into novel insights. In agriculture, machine learning and deep learning have been recently implemented to analyze images captured for various applications, such as biotic stress detection [26,27], abiotic stress detection [28,29], nitrogen estimation [30,31], spectral features selection for high-throughput phenotyping [32], weed detection [33,34] and yield prediction [17,35].…”

Retrieval of Hyperspectral Information from Multispectral Data for Perennial Ryegrass Biomass Estimation

“…As a drawback, the necessary sensor to collect such data is usually more complex and costly [ 3 ]. Furthermore, broadband absorption features may overlap and mask the discrete and informative narrowband spectral signal [ 19 ], impairing straightforward analysis and the added benefits of HS data [ 20 ]. Last, hyperspectral measurements display a high level of multicollinearity [ 21 ], providing redundant information, particularly in contiguous bands [ 22 ].…”

“…The constraints of VIs and HS are often presented as a dichotomy: while the first is seen as coarse and of limited efficiency [ 25 ], the latter is presented as highly costly and complex yet able to map subtle changes in biochemical and biophysical properties [ 20 ]. In many instances, VIs provide a partial understanding of the full spectral behavior, only directly displaying a fraction of the information contained in a spectral signature.…”