Destructive estimation of canopy chlorophyll content in smallholder crops using UAV and Sentinel-2

Loading...
Thumbnail Image

Date

Journal Title

Journal ISSN

Volume Title

Publisher

University of the Witwatersrand, Johannesburg

Abstract

Agriculture plays a crucial role in South Africa’s economy and food security, making crop health monitoring essential. Chlorophyll content is a key indicator of crop health, and its assessment can provide valuable insights into crop conditions. This study focused on a smallholder farm in the Free State Province, South Africa, aiming to estimate the chlorophyll content in wheat plants using UAV multispectral data and Sentinel-2 satellite imagery from October 2021. Vegetation indices derived from these remote sensing datasets, along with field- measured chlorophyll data obtained using a SPAD meter, were used in the analysis. Two machine learning algorithms, Support Vector Machine (SVM) and Random Forest (RF) regression, were employed to build predictive models. Four models were developed: SVM- UAV, SVM-Sentinel-2, RF-UAV, and RF-Sentinel-2. The results indicated that the RF model using Sentinel-2 data outperformed the others, achieving an R2 of 0.14 and an RMSE of 5.11 The most significant indices in this model were NDVI, GNDVI, and CIgreen. The other models produced weaker predictions: RF-UAV (R2 = 0.038, RMSE = 6.71), SVM-Sentinel-2 (R2 = 0.022, RMSE = 5.35), and SVM-UAV (R2 = 0.020, RMSE = 5.41). This study demonstrates the potential of UAV multispectral and Sentinel-2 imagery for predicting and mapping chlorophyll content in wheat crops, particularly the effectiveness of Sentinel-2 for large-scale assessments.

Description

A research report submitted in fulfillment of the requirements for the Master of Science, in the Faculty of Science, School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, Johannesburg, 2025

Citation

Rampyapedi, Napoleon. (2025). Destructive estimation of canopy chlorophyll content in smallholder crops using UAV and Sentinel-2 [Master’s dissertation, University of the Witwatersrand, Johannesburg]. WIReDSpace. https://hdl.handle.net/10539/50275

Endorsement

Review

Supplemented By

Referenced By