无人机变量施药试验设计及效果验证OA

To address the issues of excessive pesticide use and environmental pollution caused by traditional uniform spraying methods,a precision variable-rate spraying system based on drone remote sensing and deep learning intelligent identification was designed through experimental validation.The system utilizes a DJI M300 drone to capture rice field imagery,employs an optimized DETR-KD model for intelligent weed recognition and classification,integrates GIS technology to generate variable-rate spraying prescription maps,and executes precise spraying via a DJI T50 drone.Test results demonstrate that the system can reduce herbicide application by 19.4％,with key deposition parameters such as droplet coverage and sedimentation density exhibiting significant gradient-dependent responses to spraying rates,confirming the precision and pesticide reduction potential of variable-rate spraying technology.