Satellite

Random forest

Algorithm

Agriculture

Engineering

Ecology

Aerospace engineering

Biology

Statistics

Support vector machine

Agricultural engineering

Training (meteorology)

Data mining

Computer science

Regression analysis

Lookup table

Regression

Mathematics

Geography

Artificial intelligence

Remote sensing

Programming language

Meteorology

Earth observation

Artificial neural network

Machine learning

ISPRS Journal of Photogrammetry and Remote Sensing

With an upcoming unprecedented stream of imaging spectroscopy data, there is a rising need for tools and software applications exploiting the spectral possibilities to extract relevant information on an operational basis. In this study, we investigate the potential of a scientific processor designed to quantify biophysical and biochemical crop traits from spectroscopic imagery of the upcoming Environmental Mapping and Analysis Program (EnMAP) satellite. Said processor relies on a hybrid retrieval workflow executing pre-trained machine learning regression models fast and efficiently based on training data from a lookup table of synthetic vegetation spectra and their associated parameterization of the well-known radiative transfer model (RTM) PROSAIL. The established models provide spatial information about leaf area index (LAI), average leaf inclination angle (ALIA), leaf chlorophyll content (Cab) and leaf mass per area (Cm). In contrast to using site-specific training data, the approach facilitates a universal application without the need to integrate a priori information into the processor. Four machine learning algorithms, namely artificial neural networks (ANN), random forest regression (RFR), support vector machine regression (SVR), and Gaussian process regression (GPR), were found to estimate biophysical and biochemical variables of unseen targets with high performance (relative error scores < 10%). ANNs excelled in terms of accuracy, model size and execution time when the 242 spectral bands were transformed into 15 principal components, the signals of which were scaled by a z-transformation. Validation using in situ data from the SPARC03 Barrax campaign dataset revealed an overall good estimation of measured functional traits, for instance for LAI with root mean squared error (RMSE) of 0.81 m2 m−2, and for Cab RMSE of 6.2 µg cm−2 with the ANN model. Moreover, both crop traits could be successfully mapped using a pseudo-EnMAP scene revealing plausible within-field patterns. Conformity with LAI output of the SNAP biophysical processor was found especially for grassland and maize in the vegetative stages. Based on these findings, ANN models are considered the best choice for implementation of a hybrid retrieval workflow within the context of operational agricultural crop traits monitoring from future satellite imaging spectroscopy.

Efficient RTM-based training of machine learning regression algorithms to quantify biophysical &amp; biochemical traits of agricultural crops

Efficient RTM-based training of machine learning regression algorithms to quantify biophysical &amp; biochemical traits of agricultural crops

Efficient RTM-based training of machine learning regression algorithms to quantify biophysical & biochemical traits of agricultural crops