A Random Forest Algorithm for Retrieving Canopy Chlorophyll Content of Wheat and Soybean Trained with PROSAIL Simulations Using Adjusted Average Leaf Angle

Abstract

Canopy chlorophyll content (CCC) is an important indicator for crop-growth monitoring and crop productivity estimation. The hybrid method, involving the PROSAIL radiative transfer model and machine learning algorithms, has been widely applied for crop CCC retrieval. However, PROSAIL’s homogeneous canopy hypothesis limits the ability to use the PROSAIL-based CCC estimation across different crops with a row structure. In addition to leaf area index (LAI), average leaf angle (ALA) is the most important canopy structure factor in the PROSAIL model. Under the same LAI, adjustment of the ALA can make a PROSAIL simulation obtain the same canopy gap as the heterogeneous canopy at a specific observation angle. Therefore, parameterization of an adjusted ALA (ALAadj) is an optimal choice to make the PROSAIL model suitable for specific row-planted crops. This paper attempted to improve PROSAIL-based CCC retrieval for different crops, using a random forest algorithm, by introducing the prior knowledge of crop-specific ALAadj. Based on the field reflectance spectrum at nadir, leaf area index, and leaf chlorophyll content, parameterization of the ALAadj in the PROSAIL model for wheat and soybean was carried out. An algorithm integrating the random forest and PROSAIL simulations with prior ALAadj information was developed for wheat and soybean CCC retrieval. Ground-measured CCC measurements were used to validate the CCC retrieved from canopy spectra. The results showed that the ALAadj values (62 degrees for wheat; 45 degrees for soybean) that were parameterized for the PROSAIL model demonstrated good discrimination between the two crops. The proposed algorithm improved the CCC retrieval accuracy for wheat and soybean, regardless of whether continuous visible to near-infrared spectra with 50 bands (RMSE from 39.9 to 32.9 μg cm−2; R2 from 0.67 to 0.76) or discrete spectra with 13 bands (RMSE from 43.9 to 33.7 μg cm−2; R2 from 0.63 to 0.74) and nine bands (RMSE from 45.1 to 37.0 μg cm−2; R2 from 0.61 to 0.71) were used. The proposed hybrid algorithm, based on PROSAIL simulations with ALAadj, has the potential for satellite-based CCC estimation across different crop types, and it also has a good reference value for the retrieval of other crop parameters.

Growth response of black soybean (Glycine max L. Merril) with a multi canopy cropping system on measurement chlorophyll and leaf area index

“…Photoreduction of protochlorophyllide to chlorophyll a and proceeds to esterification of phytol to form a pigment chlorophyll a through the catalysis of the enzyme chlorophyllase. This change must use sunlight so that the formation of chlorophyll can occur in large quantities if the plant is exposed to a high enough IOP Publishing doi:10.1088/1755-1315/1115/1/012079 4 amount, but this amount of sunlight if it occurs too high can actually damage the chlorophyll content and enzymes that play a role in it [8,9].…”

Multi-End Physics-Informed Deep Learning for Seismic Response Estimation

“…Since traditional deep learning in a supervised manner requires a large dataset that is difficult or costly to collect, it is meaningful to train a DNN with a small dataset. Considering that the physical laws can provide additional constraints to the deep learning problem, physics-informed deep learning (PIDL, also named physics-guided or physics-constrained deep learning) has recently emerged in many areas where it takes much effort to collect a large dataset, such as computational fluid dynamics [ 15 ] and corrosion-fatigue analysis [ 16 ]. PIDL can reduce the dataset volume and simultaneously improve accuracy compared with the ordinary deep learning approaches.…”

“…Moreover, the network architectures can be designed considering the focused problems. This means we can mimic the underlying physical phenomenon using a DNN [ 31 ] or use it describe some ambiguous physical laws [ 16 , 26 , 29 ], and even combine some physics-guided computations in a NN [ 27 ]. Current PIDL studies are summarized in Table 1 .…”

“…The concept and method of applying semi-supervised learning for fully convolutional networks are introduced in detail in [ 35 ]. Since it can reduce the effort to prepare the training samples, semi-supervised learning is adopted in many engineering applications, such as building detection [ 16 ] and remaining life prognostics [ 36 , 37 ].…”

Estimation of wheat tiller density using remote sensing data and machine learning methods

“…Machine learning methods have evolved as reliable methods of learning nonlinear relationships because they require less parameterization, are implemented at various spatial and temporal scales, and are more robust and covariant to noisy features, small training sizes, and large numbers of dimensions ( Verrelst et al., 2012 ; Liang et al., 2015 ; Houborg and McCabe, 2018 ). These methods have been widely used for estimating various biophysical parameters such as the leaf area index ( Duan et al., 2019 ; Tao et al, 2020 ), vegetation cover ( Niu et al., 2021 ; Yu et al., 2021 ), biomass ( Yue et al., 2019 ; Tao et al, 2020 ), Canopy chlorophyll content ( Jiao et al., 2021 ) and the leaf tilth distribution ( Zou et al., 2022 ). However, few studies have been conducted to estimate the tiller density of winter wheat.…”