基于移动端的植株三维点云获取系统的实现

发布时间：2018-12-10 10:55

【摘要】：植株的三维模型重建以及植株的三维模型可视化是植物形态学领域的研究热点,而植株的三维点云数据是植株三维重建的基础,因此如何方便高效地获取三维点云成为了重点研究内容。目前大多数的点云获取系统都是基于计算机实现的,此类系统因安装在固定设备上,不能满足移动用户的需求。随着移动终端硬件水平的提升和无线通信网络的不断成熟,基于Android移动端的三维点云获取系统成为一种可行的解决方案。此方案能够满足人们更快捷、更方便地进行点云提取工作的需求。基于此,本文提出了一套基于移动端的植株三维点云获取方案,并依据该方案构建出了一套完整的基于移动端的植株三维点云获取系统,最后对系统进行了测试。本文主要研究内容如下:(1)提出了一种基于移动端的植株三维点云获取方案。针对PC端获取点云灵活性差和移动端处理速度慢、内存性能局限性强等问题,提出了采用客户端-服务器系统架构开发基于移动端的植株三维点云获取系统的方案。此方案旨在通过移动端获取物体的图像序列,并上传至服务器进行点云计算,最后服务器传回点云至移动端进行显示。(2)开发了基于移动端的植株三维点云获取系统。针对传统点云获取技术成本高、灵活性差等问题,采用基于移动端的植株三维点云获取方案开发了本系统。首先,从技术、理论、经济三方面对系统进行可行性分析。接着,通过对功能需求和性能需求的分析,进一步确定了系统要完成的任务。然后,通过需求分析对系统进行了总体设计和模块设计。最后,基于Android开发技术、服务器搭建技术、SIFT算法、Bundler算法和PMVS算法对系统的各个功能模块进行实现。最终开发了一套完整的基于移动端的植株三维点云获取系统。最后,为了验证本系统是否满足功能需求和性能需求,对系统进行测试。采用植株和非植株两类测试用例对系统进行了功能测试,通过控制不同变化因素对系统进行了处理时间和三维点数量两方面的性能测试。测试结果显示,本系统对于不同特征的植株和非植株物体均能较好地获取其点云数据,且当图像数量为30、图像分辨率为1600×1200时,系统处理时间保持在20分钟之内,这满足预期的功能需求和性能需求。
[Abstract]:The 3D model reconstruction and the visualization of plant 3D model are the research focus in the field of plant morphology, and the 3D point cloud data of plant is the basis of plant 3D reconstruction. Therefore, how to obtain three-dimensional point cloud conveniently and efficiently has become the focus of research. At present, most of the point cloud acquisition systems are based on computer, such systems can not meet the needs of mobile users because they are installed on fixed devices. With the improvement of hardware level of mobile terminal and the maturity of wireless communication network, 3D point cloud acquisition system based on Android mobile terminal becomes a feasible solution. This scheme can meet the needs of point cloud extraction more quickly and conveniently. Based on this, this paper proposes a plant 3D point cloud acquisition scheme based on mobile end, and constructs a complete plant 3D point cloud acquisition system based on mobile end. Finally, the system is tested. The main contents of this paper are as follows: (1) A scheme of plant 3D point cloud acquisition based on mobile end is proposed. Aiming at the problems of poor flexibility of point cloud acquisition on PC side, slow processing speed on mobile side and strong limitation of memory performance, a scheme of developing plant 3D point cloud acquisition system based on mobile end is proposed by using client-server system architecture. The purpose of this scheme is to obtain the image sequence of the object through the mobile end and upload it to the server for point cloud computing. Finally, the server sends the point cloud back to the mobile side for display. (2) A plant 3D point cloud acquisition system based on the mobile end is developed. Aiming at the problems of high cost and poor flexibility of traditional point cloud acquisition technology, this system is developed by using the scheme of plant 3D point cloud acquisition based on mobile end. Firstly, the feasibility of the system is analyzed from three aspects: technology, theory and economy. Then, through the analysis of the functional and performance requirements, the tasks to be completed by the system are further determined. Then, through the requirement analysis, the overall design and module design of the system are carried out. Finally, based on the Android development technology, server construction technology, SIFT algorithm, Bundler algorithm and PMVS algorithm to implement each functional module of the system. Finally, a complete plant 3D point cloud acquisition system based on mobile end is developed. Finally, in order to verify whether the system meets the functional requirements and performance requirements, the system is tested. Plant and non-plant test cases were used to test the function of the system. The processing time and the number of 3D points were tested by controlling different factors. The test results show that the system can obtain the point cloud data for different plant and non-plant objects, and when the number of images is 30 and the resolution of image is 1600 脳 1200, the processing time of the system is kept within 20 minutes. This meets the expected functional and performance requirements.
【学位授予单位】：西北农林科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：Q94;TP391.41

【参考文献】