Modeling and Evaluation of Biochemical Parameters in Peanut Leaves Based on Hyperspectral Data

As an important oil crop, timely and accurate estimation of biochemical parameters in peanuts is crucial for improving yield. Although hyperspectral inversion algorithms for biochemical parameters have been proven effective for a few varieties, their accuracy and applicability for high-throughput phenotyping varieties have not been fully validated. In this study, the single vegetation index method (SVI) and the random forest model (RF) were constructed to invert biochemical parameters (chlorophyll content (Cab), equivalent water thickness (Cw), and dry matter content (Cm)) and to assess the accuracy and transferability of these models across different peanut varieties. Results showed that: (1) vegetation indices of TCARI, WI and mSR were strongly correlated with Cab, Cw and Cm, respectively; (2) RF models of estimating Cab (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> = 0.77, RMSE = 8.14 μg cm<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-2</sup>), Cw (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> = 0.67, RMSE = 1.1×10<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-3</sup> g cm<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-2</sup>), and Cm (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> = 0.50, RMSE = 6.2×10<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-4</sup> g cm<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-2</sup>) demonstrated higher accuracy than the SVI models; (3) SVI models performed superior to RF models with transferability in high-throughput phenotyping varieties, with R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of 0.58 (Cab), 0.10 (Cw) and 0.19 (Cm), respectively. This study accurately estimated the biochemical parameters of peanut leaves using hyperspectral remote sensing, providing a basis for the estimation of biochemical parameters in peanuts.