Resilient Emergency Crop Harvesting Systems Under Extreme Weather: From Risk Sensing to Machinery Control and Coordinated Scheduling

Extreme rainfall, flooding, typhoons, heat and drought, early frost, and compound hazards are reshaping fundamental constraints for crop harvesting. When such hazards strike close to crop maturity, agricultural losses stem not merely from reduced field yield but propagate along a risk chain featuring degraded crop status, deteriorated field trafficability, degraded harvester performance, transport-drying congestion, and progressive grain quality deterioration. Existing investigations mostly address weather warning, remote sensing, combine-harvester manipulation, machinery scheduling, and postharvest management in isolation. Such fragmented research fails to resolve core system-level challenges for emergency harvesting, including the determination of field harvesting priority, the estimation of safe harvest duration, adaptive adjustment of machine configurations under varying crop–soil conditions, and the coordinated deployment of limited harvester fleets, transportation vehicles, and drying infrastructures. This review constructs an integrated technical chain for emergency crop harvesting. Disaster scenarios and risk formation mechanisms are analyzed; the capabilities of satellites, unmanned aerial vehicles, ground-based sensors, and crop models for retrieving maturity, lodging status, moisture content, and field operability are assessed. Adaptive chassis, headers and threshing-cleaning systems, alongside online monitoring and multi-parameter control targeting loss rate, grain breakage, impurity content, feed rate, and machine faults are comprehensively reviewed and summarized. Further analysis expands from single-machine control to field-level priority ranking, harvester-vehicle collaboration, inter-regional machinery scheduling, and the coordination spanning harvesting, drying and storage. A resilience evaluation framework oriented toward loss mitigation, operational efficiency, economic benefit, and recovery capability is proposed. Although abundant sensing, prediction, control, and optimization modules have been documented in published literature, these components are deficient in shared state variables, empirically validated risk thresholds, and closed-loop decision-making mechanisms coping with real-time perturbations. Future studies are recommended to integrate multi-source data assimilation with hybrid mechanistic-data-driven models, so as to build interpretable predictions for remaining safe harvest time, crop–soil–machine digital twins, and rolling inter-regional scheduling platforms. These technical progresses are expected to transform emergency crop harvesting from experience-driven mobilization into a quantifiable, verifiable, and restorable resilient operation system.