Study On Detection Of Characteristic Parameters Of Canopy And Variable Spray Characteristics Based On LiDAR

China ranked first in the area of orchards of the world,but the amount of pesticides used also ranks first in the world.Excessive use of pesticides not only pollutes the environment,but also reduces the quality of fruit.Reducing the use of pesticides is one of the goals of “Ane control,two reductions and three basics” put forward by the State Council of China.Improved the utilization rate of pesticides can bring significant economic and ecological such as cost saving,improving quality,emission reductions etc.The development and popularization of variable spray technology is one of the more effective ways to achieve “reduction and efficiency increase” of pesticides.In order to solve the problems of low efficiency and high environmental requirements in the orchard variable spray system,the LiDAR based detection models of Leaf Wall Area(LWA),Canopy Volume(CV),Leaf Area Density(LAD)and Leaf Area Index(LAI)have been systematically researched.The calculation model of pesticide dosage based on LiDAR and its variable spray characteristics are also studied.The main research contents and results are as follows.(1)The calculation method of calculating the reference value of canopy parameters using image processing technology is proposed,achieving the calculation of canopy's parameters reference value.It provides an accurate reference value for the conapy parameter detection model based on LiDAR.The storage method of huge point cloud data is designed,which provides technical support for LiDAR point cloud data sampling analysis and subsequent canopy parameters detection.And the FIFO storage method,designing for local point cloud data,makes it possible to improve the data processing speed of variable spray.(2)The 3D reconstruction algorithm of 2D-LiDAR point cloud data is studied.The experimental results show that the 3D reconstruction algorithm can reflect the environment size with the color depth of the point cloud.It provides a convenient way to study point cloud denoising intuitively.Denoising method of point cloud data is studied.The designed denoising method can effectively remove the useless cloud data of fruit trees.The continuous point of trunk locating approach and the proportional trunk locating approach are studied.The 3D picture shows that it is more accurate to identify the point cloud data of tree trunk by proportional tree trunk localization.(3)The canopy characteristic parameters detection model based on LIDAR is studied.The experimental results show that the detection error of LWA detection model is within 10%,which meets the requirements of the measurement accuracy.The maximum standard deviation of detection error is 3.8%,and the dispersion degree of detection results is low which indicated the model has general adaptation.The absolute error of the measurement result is basically within 20% based on the optimized CV detection model,and the measurement result is within the acceptable range.The detection errors of the LAD and LAI detection model are relatively large,mainly by the outer edge of the canopy.The leaves obscured the canopy within the canopy,resulting in the inability to completely identify all the leaves.The different scanning densities had less effect on the measurements of the four canopy characteristic parameters.(4)By studing and analyzing the method of characterizing canopy parameters with decision coefficients,a calculation model of pesticide dosage based on LiDAR point cloud data is designed.The decision coefficients of the four spraying rate calculation models are expressed by two parameters through mathematical relations.The parameters are related to the horizontal distance average,height range or horizontal distance minimum of LiDAR point cloud data.Calculate the decision coefficient K representing the characteristic parameters of the canopy in each vertical direction,and put the K into the flow formulas to obtain the partition of the flow value.(5)Using the control variable method,the multi nozzle flow model is studied.With the PID controller to rapidly stabilize the pressure and the fixed PWM frequency of solenoid valves,the multi-nozzle flow model experiment has been conducted.The results showed that the total flow rate of the pipe reached the maximum with the duty ratio of the solenoid valve PWM 65%.And through the experimental data,the formula for calculating the flow rate based on the application rate decision coefficient and the calculation formula for calculating the PWM duty cycle of the solenoid valve based on the flow rate were obtained,thus providing the basis for the flow adjustment of the variable spray system.(6)The variable spray characteristics test was carried out for the dosage calculation model.The test data showed tha the adhesion rate of the liquid spray based on the four application rate calculation models is above 95%,which is above the limit of validity of the drug liquid adhesion rate determination.Compared with continuous spraying,the rate of drug saving in the four application rate calculation models is over 68%,and the rate of drug saving is 33% or more compared with pure target spraying.LWA-based drug dosage calculation model has lower drug-saving rate,and the drug-taking rate calculation model based on CV,LAD,or LAI is higher.The above shows that under the premise of ensuring the quality of the spray,the four LiDAR-based calculation models of the application amount all achieve the purpose of saving the liquid volume.