PERTANIKA JOURNAL OF TROPICAL AGRICULTURAL SCIENCE

This study presents a comparative evaluation of manual inspection, ground-based sensors, and UAV-based remote sensing for detecting crop stress in paddy, maize, and coconut fields across Malaysia and China. Although sensing technologies have advanced considerably, cross-country comparisons between regions with differing levels of technological maturity remain limited. China, recognised for its advanced adoption of UAV and sensor-based agriculture, provides a benchmark against Malaysia’s developing digital agriculture landscape. Each method was assessed based on accuracy, responsiveness, scalability, and cost-effectiveness under field conditions. UAV-based remote sensing achieved the highest overall accuracy (mean 92%) and demonstrated superior scalability, enabling rapid large-area monitoring using vegetation indices such as NDVI and NDRE. Ground-based sensors, including soil moisture probes and chlorophyll meters, showed moderate accuracy (mean 81%) and were suitable for plot-level monitoring with real-time feedback. Manual inspection recorded the lowest accuracy (mean 68%) and limited scalability due to labour dependency and subjective assessment. UAV methods were particularly effective in early stress detection, with thermal imaging identifying canopy temperature anomalies 3–5 days before visible symptoms, especially in maize and coconut fields. Integrating UAV and ground-based sensing provided more comprehensive and timely assessments than individual approaches. These findings support the development of scalable precision agriculture frameworks tailored to tropical and subtropical systems.