Research On 3D Point Cloud Acquisition Of Plant Based On Structured Light

In order to obtain the 3D point cloud of apple tree which help build a digital apple tree model.In this paper,the 3D point cloud acquisition technology of apple tree surface is studied.A 3D point cloud acquisition method of apple tree surface relies on structured light has been proposed,the method is based on the principle of binocular stereoscopic vision.Projecting a strip laser to an apple tree what reduces the difficulty of matching the corresponding points in the two images,and the influence of the outdoor light on the strip laser can be weaken by a filter,which has special wavelength So that the 3D scanning device can be used outdoor.The main study contents are as follows:(1)A 3D point cloud acquisition device is set up,which can be used under the sunshine.The device includes a pair of ordinary network cameras,a strip laser,a stepper motors and a self-made calibration plate.It is easy to use,has low cost and simple structure.(2)According to the modified network camera,select the camera calibration algorithm,and make the calibration plate to calibrate the parameters of the camera.The calibration of the camera includes two aspects: focal length calibration and positional relationship calibration.Analyzing the calibration method and Combining with the actual situation of this study,choosing the camera displacement calibration algorithm because its calibration plate structure simple.The camera focal length is calibrated by self-made calibration plate,using three-point method for location calibration.The 3D coordinates of the calibration point on the calibration plate are calculated by the camera's focal length and the position of the two cameras.The difference between calculated distance and the actual distance of calibration point is within 0.3mm.(3)The calculation method of the 3D point is studied,which is divided into four parts: image preprocessing,line-structured light strip center extraction,stereo image matching and calculation of 3D coordinates.The method of adaptive threshold is used to extract the center of the light bar of the line structure light,and the points on the light bar in the two camera images are collocated in stereo.Finally,the straight line of the right and left ca meras and the matching points on the image are intersected to the object the actual point of the surface calculates 3D coordinates.In order to illustrate the effectiveness of this study and the scanning effect of the 3D scanning system.The indoor objects and outdoor plants were scanned and compared the point cloud data with the objects.Experiments show that the 3D scanning system proposed in this study can be used in indoor and outdoor,which can get 3D point cloud fast from common objects and trees.The accuracy of point cloud can reduce to 0.46 mm,and the average per second can get more than 2500 points,to meet the apple tree surface 3D point cloud access to the accuracy requirements.