Our PhD Students
Zubuko Njabulo Sondelani is using Multispectral imaging approach to assess drought tolerance in groundnut (Arachis hypogaea L.) genotypes and to predict yield in water stress conditions in Benin. The main objective of his study is to assess the physiological and biochemical responses of groundnut genotypes at different growth stages using multispectral imaging and predict yield under water stressed conditions. The specific objectives are to: (i) conduct a morpho-physiological characterisation of groundnut accessions collected in Benin under rainfed conditions, (ii) assess groundnut responses to drought stress based on their physiological mechanisms, (iii) identify critical growth stages of groundnuts at which water deficit stress has the most significant impact on yield using multispectral imaging based on their physiological and spectral signatures, (iv) identify key phenotypic traits associated with drought-tolerance, and (v) use machine learning models to predict the yield of groundnuts under water deficit conditions at various stages of crop development. In conclusion, the study aims to integrate stress physiology, drone imagery, digital phenotyping, and machine learning to advance drought-resilient groundnut production in Benin. Additionally, it seeks to contribute to precision agriculture approaches for climate-smart crop management.
Zubuko N. Sondelani
Natacha Quenum is working on ‘’High-through put phenotyping of okra (Abelmoschus esculentus) germplasm under drought stress using unmanned aerial vehicle (UAV)’’. Drought stress is a major constraint to okra (Abelmoschus esculentus) production in sub-Saharan Africa, where climate change has led to increasingly irregular rainfall patterns. This study aims to use transcriptomic and phenomics information for screening and identifying drought tolerant okra accessions. The specific objectives are to: (1) evaluate the effects of two irrigation regimes (30% and 100% field capacity) on the growth and yield of 320 okra cultivars under two contrasting field conditions (irrigated and managed drought); (2) identify key traits (spectral, morphological, physiological, agronomic) associated with drought stress responses in okra cultivars, ; (3) determine candidate genes linked to drought tolerance in okra; (4) develop machine learning models for predicting okra yield under drought; and (5) evaluate the interactions between genotype, environment and irrigation practices. An alpha lattice design, with two replicates per experiment will be used. During two years, agromorphological and yield data in the two research environments will be computed to generate drought tolerance indices for the selection of stable accessions. In addition, UAV-derived traits will be used to build a yield prediction model for okra under drought stress. By achieving these goals this research aims to fill the gap in phenomics selection for the development of superior okra genotypes using UAV-assisted phenotyping.
Natacha Quenum
Hermann Tiare William Some's PhD research focuses on groundnut leaf disease phenotyping using high through put approaches and computer vision models, with the goal of strengthening breeding programs in West Africa, particularly Benin. Foliar diseases, including Early Leaf Spot (ELS), Late Leaf Spot (LLS), and rust, are major constraints to groundnut productivity, yet their assessment in breeding trials remains subjective, labor-intensive, and difficult to scale.
The overarching objective is to develop an integrated AI based phenotyping framework that improves early disease detection, enables objective severity scoring, and supports data driven selection decisions. The study will (i) establish a multi modal image dataset of major groundnut leaf diseases captured at leaf, plant, and field scales using smartphones, UAVs (drones), and multispectral imagery; (ii) design computer vision models for disease identification and ordinal severity scoring aligned with standard breeder scales; (iii) develop temporal and predictive models integrating image derived traits with weather, soil, and crop growth data to enable early detection and disease forecasting; and (iv) propose a practical framework for integrating AI based phenotyping into routine breeding workflows.
Expected outcomes include validated models for automated disease scoring, deployment ready drone phenotyping tools, comparisons between AI and breeder assessments, and decision support workflows linking AI phenotypes to breeding metrics.
Hermann T. W. Some
Orthia L. F. Linkpon is working on drafting her research proposal entitled “«Optimization of Pineapple (Ananas comosus (L.) Merr.) Crop Nutrition in Ghana”. This research aims to reduce yield and quality gaps in fresh pineapple production in Ghana through precision nutrient management supported by remote sensing and machine learning. The study first conducts a yield and quality gap analysis across major pineapple production zones to quantify current performance levels and identify spatial and management-related constraints (SO1). A systematic literature review is then undertaken to assess how precision agriculture approaches can address the key nutrient-related limitations identified in high gap areas (SO2). Building on this foundation, the core of the research focuses on plant-level nutrient diagnostics. A controlled field experiment has been implemented to measure and model how nitrogen (N), phosphorus (P), and potassium (K) deficiencies alter leaf and canopy spectral signatures across multispectral bands and derived vegetation indices during the vegetative stage (SO3). Multispectral drone imagery and ground observations are used to capture the temporal spectral responses to the major nutrient stress. Finally, machine learning pipelines will be developed and validated to classify nutrient deficiency type and severity and to generate location-specific nutrient stress maps suitable for precision fertilizer recommendations (SO4). The expected outcomes include improved nutrient use efficiency, early nutrient stress detection tools, and operational decision-support products for sustainable pineapple CV. Sugarloaf production in West Africa.
Orthia L. F. Linkpon
Judicael Ganta is investigating on the topic “Effective Pest and Disease Management of Irish Potatoes (Solanum tuberosum L.) in Rwanda using multispectral imaging, artificial intelligence and agro-ecological techniques”. His research aimed to review the current state of precision agriculture technology in managing crop pests and diseases in sub-Saharan Africa, highlighting its potential benefits and challenges. It will assess Rwandan farmers’ knowledge of agro-ecological management practices for potato pests and diseases and their perceptions of factors influencing the adoption of precision farming tools. Moreover, the thesis will evaluate the effectiveness of aerial hyperspectral imagery (UAV-based remote sensing) in the early detection of bacterial wilt in potatoes. The study will also test agro-ecological management strategies for potato insects and diseases, focusing on crop associations, biopesticides, improved biofertilizers, and bio-stimulants. Lastly, the research will develop a decision support tool using convolutional neural networks to identify the severity of potato late blight, bacterial wilt, viral diseases, and insect pests and recommend appropriate treatments. This comprehensive approach will seek to enhance sustainable pest and disease management practices in Rwandan potato farming, ultimately improving crop yield and farmer livelihoods
Judicael Ganta
Tasisa Temesgen Tolossa research investigates the effectiveness of integrating Precision Agriculture (PA) and Integrated Soil Fertility Management (ISFM) to enhance soil quality and pineapple (Ananas comosus) yield in the context of climate change and declining soil fertility in Sub-Saharan Africa, particularly Ghana. The study applies chemometric analysis, machine learning, multispectral imaging, and UAV-based remote sensing to characterize soil physico-chemical properties, evaluate the impact of tillage practices, monitor spatiotemporal crop growth patterns, and predict yield using vegetation indices, soil moisture, nitrogen content, evapotranspiration, leaf area index (LAI), and chlorophyll levels. A deep learning model will also be developed for the automated detection of pineapple fruits from aerial images to improve yield estimation and farm management. Expected outcomes include improved soil quality, reduced input waste through more efficient resource management, accurate yield prediction, and data-driven decision support for farmers and agricultural stakeholders. The research will generate comprehensive datasets and analytical tools that strengthen sustainable farming practices, enhance productivity and resilience to climate impacts, inform policymakers and academic institutions, and contribute to high-quality peer-reviewed scientific publications.
Tasisa Temesgen Tolossa
Vincent Nsengimana research aims to optimize pineapple (Ananas comosus [L.] Merr.) profitability in Ghana’s Central Region by integrating artificial intelligence–based precision agriculture into advanced economic modeling frameworks. The study develops and evaluates machine learning models for automated pineapple grading and classification (Extra Class, Class I, and Class II) using image-based features, defect detection and classification (D0, D1, D3) for quality inspection, and sweetness (°Brix) prediction using robust classification and regression metrics. In parallel, the economic viability of the AI-powered grading system is assessed through a Cost–Benefit Analysis framework, while technical and profit efficiencies and their determinants are analyzed using a hybrid of Data Envelopment Analysis, Stochastic Frontier Analysis, and supervised and unsupervised machine learning approaches. Furthermore, the research examines macro-environmental drivers of industry performance using the PESTEL framework supported by multiple linear regression and machine learning models. Key expected outcomes include the development and maintenance of high-accuracy AI systems that overcome the limitations of manual and experience-based sorting, meet export and consumer quality requirements, support investment decision-making, and introduce innovative hybrid analytical methods that enhance the prediction of technical efficiency, profitability, and policy-relevant macroeconomic factors, thereby strengthening the competitiveness, sustainability, and long-term viability of the pineapple sector.
Vincent Nsengimana
Ismael Niyokwizigirwa, a PhD candidate in Crop Sciences at the University of Eswatini, is conducting research titled “Site-Specific Management of Soil Water Conservation Techniques and Organic Soil Amendments to Enhance Water Use Efficiency, Soil Fertility, and Sweet Potato Productivity in the Middleveld of Eswatini.” The study evaluates how integrated climate-smart soil and water management practices combined with organic amendments can improve sweet potato performance under variable rainfall conditions by assessing the effects of soil water conservation techniques such as mounding, ridging, and tie-ridging on soil moisture retention and yield using site-specific monitoring tools, determining the impact of organic amendments including biochar, vermicompost, and cattle manure on soil fertility and crop productivity, examining their interactive effects on water use efficiency, soil health, and yield, and analyzing how soil moisture monitoring contributes to improved tuber quality in terms of dry matter content, total soluble solids, and beta-carotene levels. Expected outcomes include improved soil moisture availability, reduced drought-related yield losses, enhanced soil fertility and moisture-holding capacity, synergistic productivity gains from integrated practices, validated use of low-cost precision tools for smallholder farmers, and contributions to climate-smart agriculture policies, extension recommendations, and sustainable sweet potato intensification strategies in Eswatini and similar semi-arid regions.
Ismael Niyokwizigirwa
Nathalie Molo research focuses on enhancing cassava productivity in Rwanda through site-specific potassium nutrient management integrated with artificial intelligence (AI) and Internet of Things (IoT)–driven precision agriculture. Addressing the challenges of high potassium demand, harvest-related nutrient losses, and soil fertility depletion associated with traditional fertilization practices, the study evaluates soil nutrient variability and potassium uptake throughout the crop cycle using laboratory analysis and handheld IoT-based soil sensors, examines the effects of different potassium fertilizer rates on cassava growth and yield across two agroecological zones in southern Rwanda, analyzes the relationship between vegetation indices and potassium response using remote sensing and AI, and develops an AI-driven biophysical model to optimize site-specific potassium application. Key expected outcomes include establishing a scientific basis for integrating portable soil sensors into soil management strategies, identifying optimal potassium rates for improved cassava performance, determining the most effective vegetation indices for nutrient monitoring, and delivering a robust AI-based model that enhances crop prediction accuracy and decision-making for precision agriculture in smallholder farming systems.
Nathalie Molo
Mandla Msimisi Makhubela is a PhD candidate in the Crop Science Department at the University of Rwanda, College of Agriculture, Forestry and Food Sciences (CAFF), School of Agriculture and Food Sciences (SAFS), whose research addresses the challenge of optimizing crop productivity by integrating advanced digital technologies into traditional farming systems. His work focuses on improving bean (Phaseolus vulgaris L.) productivity—a staple crop for millions in Rwanda and globally—through the application of Artificial Intelligence (AI) and the Internet of Things (IoT). The study introduces site-specific nitrogen nutrient management using a network of IoT soil sensors that capture real-time data on soil nutrients (nitrogen, phosphorus, and potassium), moisture, pH, electrical conductivity, and temperature, which are processed through AI-driven machine learning models to generate precise, geospatial nutrient application recommendations for farmers. The overarching objective is to develop a scalable, data-centered decision-support framework that increases productivity while reducing input costs and environmental degradation associated with uniform fertilizer application. By localizing cutting-edge precision agriculture technologies, the research directly contributes to Rwanda’s Strategic Plan for Agricultural Transformation (PSTA) and supports broader regional goals of food security and sustainable agricultural development.
Mandla M. Makhubela
Barnabas Khayil focuses on developing an integrated solar-powered smart irrigation and early fall armyworm (FAW) detection system to improve maize yields in Benin. Maize production in the region faces challenges such as inefficient irrigation, soil nutrient deficiencies, and FAW infestations, which cause significant crop loss. This research aims to combat the severe impact of Fall Armyworm (FAW) on maize, a staple crop in Benin, by developing an integrated, automated pest management system. The project will first map existing farmer practices and perceptions. It will then create a real-time detection system using thermal imaging or acoustic sensors (AudioMoth) and pair it with a hybrid solar powered, precision overhead sprinkler network for targeted pesticide application (chemigation). The core objective is to evaluate the effectiveness of this system in reducing FAW crop damage and improving maize yields relative to conventional methods.Expected key outcomes include earlier pest detection, a notable decrease in pesticide use (by 30-50%), along with lower costs and reduced labour. It also aims to minimize farmers' chemical exposure, cut down time dedicated to pest management, and boost maize yields. The study seeks to establish a scalable, sustainable model that promotes food security and aligns with global sustainability goals for smallholder farmers in Benin and other regions.
Joshua Akanson
Nathalie Molo research focuses on enhancing cassava productivity in Rwanda through site-specific potassium nutrient management integrated with artificial intelligence (AI) and Internet of Things (IoT)–driven precision agriculture. Addressing the challenges of high potassium demand, harvest-related nutrient losses, and soil fertility depletion associated with traditional fertilization practices, the study evaluates soil nutrient variability and potassium uptake throughout the crop cycle using laboratory analysis and handheld IoT-based soil sensors, examines the effects of different potassium fertilizer rates on cassava growth and yield across two agroecological zones in southern Rwanda, analyzes the relationship between vegetation indices and potassium response using remote sensing and AI, and develops an AI-driven biophysical model to optimize site-specific potassium application. Key expected outcomes include establishing a scientific basis for integrating portable soil sensors into soil management strategies, identifying optimal potassium rates for improved cassava performance, determining the most effective vegetation indices for nutrient monitoring, and delivering a robust AI-based model that enhances crop prediction accuracy and decision-making for precision agriculture in smallholder farming systems.