Predictive Modelling Employing Machine Learning, Convolutional Neural Networks (CNNs), and Smartphone RGB Images for Non-destructive Biomass Estimation of Pearl Millet

Aug 6, 2026

Upload by Eliana

Digital tools and non-destructive monitoring are essential for real-time evaluation of crop output and health in sustainable agriculture, especially for precise above-ground biomass (AGB) estimation of pearl millet. This study uses transfer learning with pre-trained convolutional neural networks alongside shallow machine learning models including support vector regression, XGBoost, and random forest regression. Smartphone RGB imaging provides cost-effective data, and Shapley additive explanations (SHAP) quantify predictor importance. SHAP identifies NGRDI and plant height as key features, with XGBoost achieving the highest accuracy (R2≈0.98; RMSE≈0.26).

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages19

Reading Time48 min

AccessFree to download

Introduction

Digital agriculture and nondestructive monitoring

Importance of biomass for crop health and carbon sequestration

Pearl millet background and resilience

Limitations of conventional destructive biomass measurement

Methods and modelling approach

Transfer learning with CNNs

Shallow machine learning models

Smartphone RGB data collection

SHAP-based feature importance analysis

Results and discussion

Most influential predictors for AGB estimation

Model comparison and performance metrics

Implications for automated monitoring and carbon inventories

TYPE Original Research PUBLISHED 06 May 2025 DOI 10.3389/fpls.2025.1594728 OPEN ACCESS EDITED BY Xing Yang, Anhui Science and Technology University, China REVIEWED BY Kavipriya J, SRM Institute of Science and Technology, India Bing Qian, Chinese Academy of Tropical Agricultural Sciences, China *CORRESPONDENCE Faten Dhawi [falmuhanna@kfu.edu.sa](falmuhanna@kfu.edu.sa)[ ](falmuhanna@kfu.edu.sa)Abdul Ghafoor [aghafoor@kfu.edu.sa](aghafoor@kfu.edu.sa) RECEIVED 20 March 2025 ACCEPTED 11 April 2025 PUBLISHED 06 May 2025 CITATION Dhawi F, Ghafoor A, Almousa N, Ali S and Alqanbar S (2025) Predictive modelling employing machine learning, convolutional neural networks (CNNs), and smartphone RGB images for non-destructive biomass estimation of pearl millet (Pennisetum glaucum) . Front. Plant Sci. 16:1594728 . doi: 10.3389/fpls.2025.1594728 COPYRIGHT © 2025 Dhawi, Ghafoor, Almousa, Ali and Alqanbar. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Predictive modelling employing machine learning, convolutional neural networks (CNNs), and smartphone RGB images for non-destructive biomass estimation of pearl millet (Pennisetum glaucum) Faten Dhawi 1*, Abdul Ghafoor 2*, Norah Almousa 3, Sakinah Ali 3 and Sara Alqanbar 3 1Agricultural Biotechnology Department, College of Agricultural and Food Sciences, King Faisal University, Al Ahsa, Saudi Arabia, 2Center for Water and Environmental Studies, King Faisal University, Al-Ahsa, Saudi Arabia, 3 Fab Lab, Abdulmonem Al Rashed Humanitarian Foundation, Al-Ahsa, Saudi Arabia Digital tools and non-destructive monitoring techniques are crucial for real-time evaluations of crop output and health in sustainable agriculture, particularly for precise above-ground biomass (AGB) computation in pearl millet (Pennisetum glaucum) . This study employed a transfer learning approach using pre-trained convolutional neural networks (CNNs) alongside shallow machine learning algorithms (Support Vector Regression, XGBoost, Random Forest Regression) to estimate AGB. Smartphone-based RGB imaging was used for data collection, and Shapley additive explanations (SHAP) methodology evaluated predictor importance. The SHAP analysis identiﬁed Normalized Green-Red Difference Index (NGRDI) and plant height as the most inﬂuential features for AGB estimation. XGBoost achieved the highest accuracy (R2 = 0 . 98, RMSE = 0 . 26) with a comprehensive feature set, while CNN-based models also showed strong predictive ability. Random Forest Regression performed best with the two most important features, whereas Support Vector Regression was the least effective. These ﬁndings demonstrate the effectiveness of CNNs and shallow machine learning for non-invasive AGB estimation using cost-effective RGB imagery, supporting automated biomass prediction and real-time plant growth monitoring. This approach can aid small-scale carbon inventories in smallholder agricultural systems, contributing to climate-resilient strategies. KEYWORDS digital agriculture, deep learning, plant monitoring, carbon sequestration, CNN convolutional neural network Frontiers in Plant Science 01 [frontiersin.org](frontiersin.org) 1 Introduction In sustainable agriculture, digital technologies and nondestructive monitoring methods are essential for real-time assessments оf crop health and productivity. Biomass serves as an indicator оf crop vigor, reﬂecting vital processes such as photosynthesis, energy transfer, and water exchange with the atmosphere (Yue et al., 2017; Liu et al., 2022; Zheng et al., 2023) . Accurat

FAQs

How does the study estimate above-ground biomass (AGB) without destroying plants?

It combines smartphone-based RGB imaging with machine learning, including transfer learning CNNs and shallow models, to predict AGB from image-derived predictors.

Which features were found to be most important for AGB estimation?

The SHAP analysis identified the Normalized Green-Red Difference Index (NGRDI) and plant height as the most influential features.

Which model achieved the best predictive performance and what were its metrics?

XGBoost produced the highest accuracy, with R2 around 0.98 and RMSE around 0.26 when using a comprehensive feature set.

Aug 6, 2026

Upload by Eliana