Sub-Saharan Africa’s cereal yields remain ~60% below global benchmarks, while unmanned aerial vehicle (UAV) adoption in smallholder systems averages below 2–3% across major economies, revealing a performance–adoption disconnect that requires systematic investigation. This Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 synthesis of 127 sources quantifies the performance of UAV sensors and identifies mechanisms that constrain their adoption across regional agricultural systems. Random-effects meta-analysis synthesized evidence from 81 quantitative studies, yielding 101 total observations. Pooled detection accuracy was estimated from 49 studies contributing 52 observations (mean 90.2%, 95% confidence interval (CI): 89.8–92.6%). Yield prediction performance was assessed from 32 studies contributing 49 observations (pooled coefficient of determination (R2) = 0.841, 95% CI: 0.827–0.855), validating technical feasibility. Cost-effectiveness analysis reveals significant performance–price differentiation: red-green-blue (RGB) sensors achieve 89.4% accuracy at United States Dollar (USD) 16.50 per percentage point versus hyperspectral systems at 93.7% accuracy but at USD 132.17 per point, resulting in a 25.6 times cost differential. Yield prediction models demonstrate robust performance (R2 = 0.81; cereal crops R2 = 0.82). Barrier analysis identifies economic constraints as the primary limiter, with capital requirements reaching 0.8–3.1 times the annual smallholder income. Infrastructure deficits impose secondary constraints, particularly in rural electrification, below 50%. Case study synthesis reveals that coordinated interventions addressing multiple barriers simultaneously—cooperative ownership, off-grid infrastructure, and streamlined regulation—achieve substantially higher adoption than isolated approaches. Engineering economics positions RGB platforms for individual deployment and multispectral systems for cooperative scales (20–50 farmers), establishing feasible deployment pathways for tens of million regional smallholder operations.
Akande et al. (2026) studied this question.