玉米三维模型的重构及时序插值技术研究
发布时间：2019-01-10 14:30
【摘要】：随着科技的进步,计算机图形学与虚拟技术得到很大的发展,利用计算机建立虚拟模型也成为一个热门话题。日常生活中所见到的植物不仅种类繁多,更为整个世界增添了很多色彩,为人们提供了赖以生存的食物。因此,植物虚拟三维模型的建立已成为图形学领域的一大热点,而且虚拟植物模型的应用前景非常广泛,比如设计数字化城市、制作游戏场景以及进行遥感实验。对农作物进行虚拟模型的建立可以进行虚拟试验以研究出更高产量的农作物,建立虚拟农场可以更好地实现农田的管理。本文主要以玉米为目标进行虚拟模型研究,主要研究内容有以下几个方面:1、几何参数建模方法研究。对非均匀B样条曲面(简称NURBS)的概念进行阐述,了解NURBS曲面拟合所需条件:控制点和节点矢量。对玉米植株进行测量获得地上各部分器官的几何参数,然后根据控制点相关算法求取控制叶鞘和叶片组成的NURBS曲面的控制点和节点矢量,再根据NURBS拟合算法得到表示叶鞘和叶片的NURBS曲面。根据得到的茎秆高和半径,画圆柱代替玉米的茎秆,最后将两个模型组合起来得到完整玉米植株的模型。2、点云数据建模方法研究。从三个方向对玉米植株进行扫描获得玉米植株表面点的三维空间坐标,再对获得的数据进行拼接得到完整玉米植株数据。由于数据量巨大,首先对数据进行一些预处理,如去噪、精简等;另外,玉米的叶片类似于二维空间的一个曲面,而茎秆相当于圆柱,为了方便建模,对叶片和茎秆进行分割;然后利用区域增长算法实现叶片建模,并用圆柱拟合算法实现茎秆的建模;最后将得到的模型整合在一起组成完整的植株模型。3、时序插值方法研究。测量两个时段玉米植株的几何参数,然后利用关键帧技术的插值函数求出两个时段几何参数之间的线性关系,最后利用这个线性关系得到玉米中间状态的生长过程并得到其模型。
[Abstract]:With the progress of science and technology, computer graphics and virtual technology have been greatly developed, the use of computers to build virtual models has become a hot topic. The plants seen in daily life are not only diverse, but also the whole world has added many colors, providing people with food for survival. Therefore, the establishment of virtual 3D models of plants has become a hot topic in the field of graphics, and the application prospect of virtual plant models is very extensive, such as designing digital cities, making game scenes and remote sensing experiments. Virtual experiment can be carried out to study the crops with higher yield, and the management of farmland can be realized better by establishing virtual farm. The main contents of this paper are as follows: 1. Geometric parameter modeling method. The concept of nonuniform B-spline surface (NURBS) is introduced, and the conditions of NURBS surface fitting are obtained: control point and node vector. The geometric parameters of the organs of each part of the ground were measured and the control points and nodal vectors of the NURBS surface composed of the leaf sheath and the leaf were obtained according to the control point correlation algorithm. Based on the NURBS fitting algorithm, the NURBS surfaces representing the sheaths and blades are obtained. According to the height and radius of the stem, the stem of maize was replaced by a cylinder. Finally, the two models were combined to get the model of the whole maize plant. 2. The method of point cloud data modeling was studied. The three dimensional coordinates of the surface points of maize plants were obtained by scanning the maize plants in three directions, and the complete data of maize plants were obtained by splicing the obtained data. Because of the huge amount of data, the data is preprocessed first, such as de-noising, streamlining and so on. In addition, the leaves of maize are similar to a curved surface in two-dimensional space, and the stem is equivalent to a cylinder. In order to facilitate modeling, the leaves and stems are divided. Then the region growth algorithm is used to model the leaves, and the cylindrical fitting algorithm is used to model the stem. Finally, the obtained model is integrated to form a complete plant model. 3, the time series interpolation method is studied. The geometric parameters of corn plants in two periods were measured, and the linear relationship between the geometric parameters of two periods was obtained by using the interpolation function of key frame technique. Finally, the growth process of corn intermediate state is obtained by using this linear relation and its model is obtained.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：S126;S513
本文编号：2406421
[Abstract]:With the progress of science and technology, computer graphics and virtual technology have been greatly developed, the use of computers to build virtual models has become a hot topic. The plants seen in daily life are not only diverse, but also the whole world has added many colors, providing people with food for survival. Therefore, the establishment of virtual 3D models of plants has become a hot topic in the field of graphics, and the application prospect of virtual plant models is very extensive, such as designing digital cities, making game scenes and remote sensing experiments. Virtual experiment can be carried out to study the crops with higher yield, and the management of farmland can be realized better by establishing virtual farm. The main contents of this paper are as follows: 1. Geometric parameter modeling method. The concept of nonuniform B-spline surface (NURBS) is introduced, and the conditions of NURBS surface fitting are obtained: control point and node vector. The geometric parameters of the organs of each part of the ground were measured and the control points and nodal vectors of the NURBS surface composed of the leaf sheath and the leaf were obtained according to the control point correlation algorithm. Based on the NURBS fitting algorithm, the NURBS surfaces representing the sheaths and blades are obtained. According to the height and radius of the stem, the stem of maize was replaced by a cylinder. Finally, the two models were combined to get the model of the whole maize plant. 2. The method of point cloud data modeling was studied. The three dimensional coordinates of the surface points of maize plants were obtained by scanning the maize plants in three directions, and the complete data of maize plants were obtained by splicing the obtained data. Because of the huge amount of data, the data is preprocessed first, such as de-noising, streamlining and so on. In addition, the leaves of maize are similar to a curved surface in two-dimensional space, and the stem is equivalent to a cylinder. In order to facilitate modeling, the leaves and stems are divided. Then the region growth algorithm is used to model the leaves, and the cylindrical fitting algorithm is used to model the stem. Finally, the obtained model is integrated to form a complete plant model. 3, the time series interpolation method is studied. The geometric parameters of corn plants in two periods were measured, and the linear relationship between the geometric parameters of two periods was obtained by using the interpolation function of key frame technique. Finally, the growth process of corn intermediate state is obtained by using this linear relation and its model is obtained.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：S126;S513
【参考文献】
相关期刊论文 前2条
1 严涛,陈彦云,吴恩华;一种基于单幅图像的树木深度估计与造型方法[J];计算机学报;2000年04期
2 唐卫东;李金忠;刘昌鑫;胡雪华;李萍萍;卢章平;;虚拟植物模型及其构建方法研究综述[J];计算机应用研究;2012年09期
，本文编号：2406421
本文链接：https://www.wllwen.com/kejilunwen/nykj/2406421.html
