From Wide-Area Screening to Precise Diagnosis: A Two-Step Air-Ground Collaborative Approach for the Detection of Citrus Huanglongbing

In response to the urgent need for efficient, precise and low-cost monitoring of Huanglongbing in large-scale citrus orchards, this study explored a two-step technical framework of “wide-area screening to precise diagnosis” for the aerial-ground collaboration. In the wide-area screening stage of the drone, a lightweight YOLOv8-SNNL new model was constructed, which significantly reduced the number of parameters while enhancing the ability to capture subtle symptoms of the canopy. This model achieved an accuracy rate of 90.6%, a recall rate of 93.3%, and an mAP50 of 96.6% on the independent test set, enabling efficient and reliable positioning of suspected diseased plants. Then, the ground operators reached the corresponding locations based on the geographic coordinate position information output by the drone. Using the constructed SRSA-YOLO-World model for ground precise diagnosis, the constructed model achieved an identification accuracy of 99.5% for diseased leaves in complex field environments. Based on the positioning and diagnosis results output by both models, managers were provided with a decision-making basis. The aerial-ground collaboration strategy integrates the efficiency of drone “wide-area screening” and the precision of ground equipment “precise diagnosis”, providing a new solution for replacing the traditional manual inspection mode and achieving precise prevention and control in large-scale orchards.