Research On Plant Height Measurement Method Based On Lidar

Plant height is one of the indicators to measure the growth rate of plants,and it is a main aspect of the determination of its growth status.Regular measurement of plant height is of great significance for understanding the relationship between plant growth rate and external meteorological conditions.This research is based on the Agricultural Meteorological Automated Observation Station project of the Newsky Co.,Ltd.of the 23 Institute of China Aerospace Science and Industry Corporation Limited.The project is developed according to the requirements of the "Crop Meteorological Automated Observation Station Functional Specification" of the Comprehensive Observation Department of China Meteorological Administration.Plant height is an integral part of the project.The purpose of this paper is to use two-dimensional laser radar(LiDAR)to extract plant height information of plants.According to the scanning method and data acquisition characteristics of the two-dimensional LiDAR,a LiDAR elevation scanning device is designed.The device is mainly composed of two servo motors and a two-dimensional LiDAR.The servo motor is used to control the LiDAR.Pitch angle,add a scanning degree of freedom through rotation,modify the format of the radar output data,add a LiDAR pitch angle value,and output the data in the form of text.The device can complete the task of automatic plant scanning.The data processing of LiDAR includes the coordinate conversion of the original radar data,the segmentation of the canopy data and the correction of the data.Convert the two-dimensional LiDAR data into a three-dimensional rectangular coordinate system.The converted three-dimensional point cloud model is divided into ground point cloud and canopy point cloud,but the overall point cloud has an overall tilt,and the segmented ground point cloud needs to be segmented Use Random Sampling Consistency Algorithm(RANSAC)to get the fitting straight line,and the inclination angle obtained from the fitting straight line,so as to correct the overall point cloud model.Filter the plant canopy point cloud data.The filtering method of multi-plant canopy point cloud is analyzed in detail.Based on the distance and distribution of the point cloud,a curve detection-normal kernel density algorithm is proposed.The three-dimensional point cloud is projected to a two-dimensional plane,and the projection point is used to fit the canopy.The distance distribution law of the curve filters out outliers.The plant height is calculated by the height of the treated canopy to the ground.Comparing the LiDAR and manual measurement data,it is concluded that the plant height measurement method based on LiDAR has good accuracy and stability.A plant data processing system is designed using Matlab software,and the processing algorithms used in the above research content are integrated into the system,and the processed point cloud image is automatically output,and the height data of the plant is calculated.Compared with the existing measurement methods,the plant height measurement research method has the advantages of simple operation,low measurement environment requirements,and high cost performance..