1. Introduction

Maize is the primary staple crop in Malawi and is predominantly cultivated by smallholder farmers operating on plots typically smaller than one hectare. Smallholder producers of this kind account for the majority of agricultural output across Sub-Saharan Africa (FAO, 2013). Accurate and timely yield estimation is therefore essential for national food security planning, early warning systems, and resource allocation. However, traditional field-based yield assessment methods are labour-intensive, time-consuming, costly, and difficult to scale across dispersed rural farming systems. Survey-based agricultural statistics in the region are further constrained by measurement error and limited spatial coverage, which reduces their reliability for planning purposes (Carletto et al., 2019).

With the increasing availability of open-source geospatial tools and freely accessible satellite imagery, remote sensing offers a scalable and cost-effective alternative for crop yield monitoring in low-resource settings. Reviews of multi-sensor approaches indicate that optical, radar, and hyperspectral data each provide distinct advantages for crop monitoring depending on the agricultural setting (Ali et al., 2022). Satellite-based methods have already been applied to maize yield estimation in intercropped smallholder fields in Southern Malawi, demonstrating the feasibility of medium-resolution imagery in this context (Li et al., 2022).

This research nevertheless addresses a persistent gap in the literature. While previous work has evaluated satellite data for maize yield estimation in Malawi (Li et al., 2022), no prior study has systematically compared UAV, Sentinel-2, and Landsat imagery within a single smallholder framework. Existing research has largely examined single-sensor approaches, or has been conducted in agricultural systems with substantially larger field sizes, where medium-resolution satellite pixels align more readily with field boundaries.

This study therefore evaluates and compares the performance of Sentinel-2, Landsat 9, and Unmanned Aerial Vehicle (UAV) imagery for maize yield prediction in selected smallholder fields in Zomba District, Malawi. The objective is to assess prediction accuracy across sensors and to demonstrate how open-source geospatial technologies can support scalable yield monitoring in Sub-Saharan Africa.

1. Materials and Methods

2.1 Study area and ground-truth data

The study area consisted of maize fields ranging from 0.06 to 0.89 hectares, representative of Malawi's smallholder farming systems. Ground-truth yield data were collected at the end of the 2024 to 2025 growing season through manual harvesting and weighing of maize cobs. Yields were standardized to tons per hectare to enable comparison across fields of varying sizes.

2.2 Image acquisition

2.2.1 Satellite imagery
Imagery from Sentinel-2 was accessed through Google Earth Engine and acquired monthly from November to March. Landsat 9 imagery was downloaded from the USGS Earth Explorer for the same period, excluding January due to persistent cloud cover.

2.2.2 UAV imagery
UAV data were collected in March 2025 using a DJI Matrice 300 RTK flown at an altitude of 120 metres, achieving a ground sampling distance of 3.67 cm per pixel. The UAV was equipped with a Zenmuse H20N camera. Drone imagery was processed into orthomosaics using Pix4DMapper before further analysis in QGIS.

2.3 Image processing and feature extraction
Image processing and feature extraction were conducted using QGIS, emphasizing the use of open-source tools to ensure accessibility and reproducibility. Vegetation indices derived from reflectance in the visible and near-infrared regions are widely established as proxies for canopy biomass and vigour, and form the basis of most remote sensing approaches to crop monitoring (Ahamed et al., 2011).

2.3.1 Satellite features
Vegetation indices including NDVI, SAVI, and GNDVI were computed, along with extraction of raw Red, Green, and Blue spectral bands. Monthly averages were calculated at the field level. To increase the number of training samples, each monthly satellite image was treated as an independent observation.

2.3.2 UAV features
High-resolution orthomosaics enabled extraction of RGB-based vegetation indices including Excess Green (ExG), Green Leaf Index (GLI), Visible Atmospherically Resistant Index (VARI), Triangular Greenness Index (TGI), and Colour Index of Vegetation Extraction (CIVE), along with mean RGB values. Although UAV imagery was collected only once during crop maturity, its high spatial resolution provided detailed canopy information.

2.4 Model development and evaluation
Separate linear regression models were developed for each sensor type using the extracted indices and band features as predictors. Linear regression was selected due to its simplicity, interpretability, and suitability for small datasets, particularly in contexts where field sizes are small and the number of usable pixels is limited. This constraint is most acute for coarse-resolution imagery such as Landsat 9, which has a spatial resolution of 30 metres. More complex machine learning models risk overfitting and may not generalize well in smallholder farming environments.

Model performance was evaluated using the Coefficient of Determination (R²), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE). For UAV data, cross-validation was conducted by rotating through fields to ensure robust evaluation.

1. Results

Results revealed substantial differences in predictive performance across sensors. UAV imagery achieved the highest accuracy, with an R² of 0.9991, RMSE of 0.0135 t/ha, and MAE of 0.0109 t/ha, indicating an almost perfect fit between predicted and observed yields. Sentinel-2 demonstrated moderate predictive capability, achieving an R² of 0.4063, RMSE of 0.3537 t/ha, and MAE of 0.2942 t/ha. Landsat 9 exhibited the lowest performance, with an R² of 0.2331, RMSE of 0.4020 t/ha, and MAE of 0.3365 t/ha. A summary of model performance across the three sensors is presented in Table 2.

These findings suggest that while UAV imagery provides highly precise yield estimates due to its fine spatial resolution, satellite-based approaches, and particularly Sentinel-2, offer promising scalability despite lower predictive strength in this dataset. The satellite results are broadly consistent with the accuracy ranges previously reported for smallholder maize systems in Southern Malawi (Li et al., 2022).

1. Discussion
   The superior performance of UAV imagery can be attributed to its high spatial resolution, which captures within-field variability more effectively than medium-resolution satellite data. However, UAV deployment involves higher operational costs, logistical coordination, and technical expertise, all of which limit its scalability for nationwide monitoring. In contrast, Sentinel-2 and Landsat 9 imagery are freely available, offer broad spatial coverage, and support continuous time-series analysis, making them more practical for large-scale agricultural monitoring in low-resource environments. Improved performance may be achieved by expanding training datasets, incorporating longer time series, or integrating multi-sensor data fusion approaches, the potential of which has been noted in comparative reviews of sensor combinations (Ali et al., 2022).

Reliable field-level yield information also underpins agronomic decision-making at the farm scale. Returns to inputs such as fertiliser vary considerably across smallholder plots (Duflo et al., 2008), and uncertainty about expected outcomes strongly influences whether farmers adopt improved technologies (Emerick et al., 2016). Yield estimates that can be produced repeatedly and at low cost therefore have value beyond national statistics, informing targeted advisory services and input allocation.

A key contribution of this study is its demonstration of a fully open and reproducible workflow using QGIS and freely available imagery. By leveraging open-source software and public satellite data, the framework reduces dependency on proprietary systems and enables local researchers, agricultural officers, and non-governmental organizations to implement crop monitoring systems without costly infrastructure. This approach strengthens local capacity for geospatial analysis and supports sustainable agricultural monitoring initiatives in Sub-Saharan Africa, where the quality of conventional agricultural statistics remains a recognised constraint (Carletto et al., 2019).

The findings highlight the trade-offs between spatial resolution, cost, scalability, and predictive accuracy. While UAV imagery achieved the highest precision, Sentinel-2 demonstrated potential as a scalable alternative for early yield estimation when combined with appropriate modelling approaches and larger datasets.

1. Conclusions

The integration of open data, simple and interpretable models, and open-source geospatial tools offers a viable pathway for affordable and repeatable crop yield estimation in smallholder systems. With further refinement and expanded datasets, satellite-based approaches can play a significant role in supporting food security decision-making across Malawi and similar Sub-Saharan African contexts, complementing rather than replacing the conventional survey instruments on which national statistics currently depend (FAO, 2013). The methods and processing workflows developed in this study are openly shared to facilitate reuse and adaptation by researchers and practitioners aiming to strengthen agricultural monitoring and planning efforts.

References
Ahamed, T., Tian, L., Zhang, Y., Ting, K. C., 2011. A review of remote sensing methods for biomass feedstock production. Biomass and Bioenergy, 35(7), pp. 2455–2469. https://doi.org/10.1016/j.biombioe.2011.02.028
Ali, A. M., Abouelghar, M., Belal, A. A., et al., 2022. Crop yield prediction using multi sensors remote sensing (review article). Egyptian Journal of Remote Sensing and Space Sciences, 25(3), pp. 711–716. https://doi.org/10.1016/j.ejrs.2022.04.006
Carletto, C., Yacoubou Djima, I., Tiberti, M., 2019. Improving the availability, quality, and policy relevance of agricultural statistics in Mali. USAID & World Bank. https://pdf.usaid.gov/pdf_docs/PA00XR99.pdf (accessed 22 July 2026)
Duflo, E., Kremer, M., Robinson, J., 2008. How high are rates of return to fertilizer? Evidence from field experiments in Kenya. American Economic Review, 98(2), pp. 482–488. https://doi.org/10.1257/aer.98.2.482
Emerick, K., de Janvry, A., Sadoulet, E., Dar, M. H., 2016. Technological innovations, downside risk, and the modernization of agriculture. American Economic Review, 106(6), pp. 1537–1561. https://doi.org/10.1257/aer.20150474
FAO, 2013. Smallholders and family farmers. Food and Agriculture Organization of the United Nations. https://www.fao.org/family-farming/detail/en/c/273864/ (accessed 22 July 2026)
Li, C., Chimimba, E. G., Kambombe, O., Brown, L. A., Chibarabada, T. P., Lu, Y., Anghileri, D., Ngongondo, C., Sheffield, J., Dash, J., 2022. Maize yield estimation in intercropped smallholder fields using satellite data in Southern Malawi. Remote Sensing, 14(10), 2458. https://doi.org/10.3390/rs14102458