Research On Reconstruction And Temporal Interpolation Technology Of 3D Corn Model

With the progress of the science and technology, the computer graphics and virtual technique has been greatly developed, therefore, using computer to create a virtual model has become a hot topic. A wide variety of plants seen in the daily life not only make the whole world colorful, but also provide necessary food to survive. Therefore, the establishment of three-dimensional virtual plant model has become a hot area of the computer graphics, whose application prospect is very extensive, such as the design of the digital cities, manufacture of the game scene and experiment of the remote sensing. The create of crops virtual model can perform virtual experiments to study a higher yield of crops, and the establishment of virtual farms can better realize farmland management.In this thesis, using corn as a target to study the virtual plant model, the main study contents are as follows:1. The study on the geometric parameters of the modeling method. The concept of non-uniform rational B-splines surfaces(NURBS surfaces) is elaborated for understanding the required conditions of the NURBS surface fitting(the control points and the knot vector). The corn plant is measured to obtain the geometric parameters of each section of corn organs. Then, based on the control point correlation algorithm, find the control points and node vectors of the NURBS surfaces which consist of the leaf sheaths and leaf blades. Thus, the leaf sheaths and blades can be obtained according to the NURBS surface fitting algorithm. Next, draw the cylinder to represent the corn stalk according to the obtained height and radius of the stalk. Finally, the abovementioned two models are combined to get a complete model of the corn plant.2. The study on point cloud data of the modeling method. Scan corn plant from three directions to obtain the three-dimensional coordinate of corn plant surface point, and then get the data of the whole corn plant. Due to the enormous amount of data, the data pretreatment is made first, such as denoising, streamlining and so on. In addition, make the leaf blade of the corn as a two-dimensional surface and the stalk as a cylinder to facilitate the modeling of the leave and stalk split. Then use the regional growth algorithm to achieve blade modeling and cylinder fitting algorithm to achieve stalk modeling. Finally, combine the obtained modelings to get a whole plant model.3. The study on time domain interpolation method. Measure the geometric parameters of the corn plant in two different periods. Then, use the interpolation function of the key frame technology to calculate the linear relationship of the geometric parameters between two periods. Finally, the growth progress of the corn state and its model are obtained using this linear relationship.