Development Of Three-dimensianl Shape Information Acquisition System For Plant Leaves

Leaf blade is an important part of plants.Accurately Obtaining the three-dimensional shape information of the blade can be used for visual simulation of plant leaves and functional analysis of leaves.It is of great significance for the study of virtual plants,and can promote the combination of agriculture and computer technology。At the same time,the information can help the development of agricultural informationization.At present,the main methods for three-dimensional measurement of plant leaves are mainly classified into contact type and non-contact type,and the method based on structured light is generally considered to be practical three-dimensional information acquisition method.With the update of technology,the structure light measurement method has made new progress,and the accuracy of data acquisition is gradually improving.In order to obtain the threedimensional shape information of the blade in real time at low cost,a system based on the principle of structured light measurement was designed,and the structural light processing method which is suitable for the blade was selected.The main work of this thesis is as follows:(1)Analyzed the imaging principle of the camera and clarify the content of the camera calibration.According to Zhang Zhengyou’s camera calibration method,combined with OpenCV to achieve the calibration function of the system,after experimentation,the calibration of the system can achieve a certain degree of precision.(2)For the three-dimensional measurement of plant leaves,the measurement model of the line structure light used was studied in detail,and the solution of the model was analyzed to realize the mathematical transformation of the two-dimensional point in the image to the three-dimensional point in the camera coordinate system.According to the operating efficiency of the solution,the rapid solution method was selected for the system.(3)In the aspect of calibration of structured light parameters,ie optical plane calibration,based on the perspective of the perspective projection geometry of the camera,a light plane calibration algorithm based on checkerboard was designed.The algorithm does not need to solve the camera external parameters every time,the calculation is relatively simple,.And the plane fitting based on RANSAC principle was used to effectively reduce the influence of the error strip extraction point on the light plane parameters.(4)A detailed study of common light strip center extraction algorithm.The extremum method and the threshold method was combined when extracting the light strip by the gray center of gravity method,and the appropriate threshold parameter was selected according to the characteristics of the blade.At the same time,the Hessian matrix method and the direction template method were studied.After optimizing the suitable parameter conditions,the contrast experiment of light strip extraction was carried out.The strip center extraction algorithm for the system is selected by running speed and extraction effect.After all the algorithms have been tested and selected,the system’s visual operation interface was realized based on Qt,the system software development was completed finally.Experiments were carried out to obtain 3D point cloud information of plant leaves by using the software.Through the evaluation of the distance measurement,the accuracy can reach 99%.