Coordinated Regulation of Orchard Variable-Rate Spraying: From Canopy Perception to Actuation—A Review

Orchard variable-rate spraying increasingly depends on the coordinated conversion of heterogeneous perception outputs into spatially and temporally matched actuator commands, yet the evidence linking sensing improvements to measurable spray-performance gains remains fragmented. Rather than reviewing canopy perception technologies in isolation, this review focuses on the complete perception-to-actuation chain for coordinated orchard spraying and identifies which perception outputs are sufficiently actionable for regulation of liquid flow, airflow, nozzle timing, and spray trajectory. A structured narrative review of the peer-reviewed literature published up to August 2026 was organized by perception object, sensing modality, platform type, orchard system, and control relevance; representative studies were retained when they reported a clear linkage between sensed information and spraying decisions or provided transferable evidence for that linkage. Four perception-information classes are distinguished: canopy geometric profile, internal canopy structure, crop/plant-protection status, and spatial-geographic/operational context. LiDAR, RGB/RGB-D and binocular vision, multispectral/hyperspectral sensing, positioning, and multimodal fusion are compared according to the information they provide and their suitability for real-time control. Cross-study synthesis shows that the strongest evidence currently concerns canopy-volume- or density-based liquid-flow-rate regulation, while robust mappings for airflow demand, disease-severity-based dosage, and cross-season transfer remain less mature. Reported benefits include substantial pesticide savings in selected field or prototype studies, but the evidence is heterogeneous and often species-, platform-, and scenario-specific. The central research need is, therefore, not higher perception accuracy alone, but physically interpretable, latency-aware, spatially registered mappings from heterogeneous perception outputs to executable nozzle, liquid-flow, airflow, and trajectory commands. Concrete priorities include standardized structural descriptors, end-to-end delay compensation, transferable dose models, and benchmark field protocols for closed-loop validation.