【摘要】 果树冠层是果树最先接触大气与光照的部位，也是果树进行光合作用和呼吸的直接通道，与水果的品质、产量息息相关。为了提高农药的利用率，论文利用激光雷达与双目视觉传感器获取果树冠层的特征数据并进行数据融合，用以指导变量喷雾。通过激光雷达获取果树冠层点云数据并对点云数据进行分隔、滤波、匹配等处理后，再对冠层进行重构，从上至下将冠层分为四层，计算每层的相应密度占比；通过双目视觉传感器结合2G-R-B算法提取果树冠层形态数据，计算出叶墙的像素面积，根据参照物对叶墙面积进行标定，进而计算出四部分叶墙面积(LWA)的真实大小。喷雾系统根据每层的冠层密度、叶墙面积数据，通过PWM控制电磁阀的工作状态实现流量实时可调，完成果树冠层的变量喷雾。与常规连续喷雾相比，变量喷雾的省药率约为21.02%,减少了药液的浪费。更多还原

【Abstract】 Fruit tree canopy is the first part of fruit trees to contact with the atmosphere and light, and is also the direct channel for photosynthesis and respiration of fruit trees, which is closely related to the quality and yield of fruit.In order to improve the utilization rate of pesticides, LiDAR and binocular vision sensor are used to obtain characteristic data of fruit tree canopy and data fusion to guide variable spray.After obtaining the point cloud data of fruit tree canopy by LiDAR,filtering and matching the point cloud data, the canopy is reconstructed and divided into four layers from top to bottom, and the corresponding density proportion of each layer is calculated.The canopy shape data of fruit trees are extracted through binocular vision sensor combined with 2G-R-B algorithm, and the pixel area of leaf wall is calculated.The leaf wall area is calibrated according to the reference object, and the real size of four parts of leaf wall area(LWA) is calculated.According to the data of canopy density and leaf area of each layer, the spray system is based on the canopy density and leaf area data of each layer and the working state of the solenoid valve is controlled by PWM to achieve a real-time adjustable flow rate to complete the variable spray in the canopy of the fruit tree.Compared with conventional continuous spray, the rate of variable spray saves approximately 21.02 %,and the waste of liquid is reduced.更多还原