基于目标识别和参数化技术的城市建筑群三维重建研究

发布时间：2018-03-31 06:18

本文选题：城市建筑群　切入点：三维重建　出处：《浙江大学》2013年博士论文

【摘要】：随着数字城市的快速发展和城市三维空间模型应用领域的不断扩展,城市建筑群三维模型的需求日益增长。然而传统建筑物三维重建方法在面对大空间尺度、大数据量、更新节奏快的城市建筑群时,在效率、精度、成本、尺度、技术门槛等方面均不同程度存在缺陷。为此,论文尝试引入目标识别和参数化技术,探求一种适用于城市建筑群的低成本、低门槛、高效率的“大众化”三维重建解决方案,以满足数字城市及相关领域的迫切需求。论文的主要内容包括：(1)对建筑物三维重建、目标识别和参数化建模这三大领域的相关研究、技术方法作了总结回顾,指出了其存在的问题；(2)对三大技术体系作了深度解构,指出了“二元并行框架”的成因,给出了交叉的可行途径,并通过体系重构,构建了城市建筑群三维重建的“三元交叉框架”；(3)针对遥感影像分割,提出了面向对象的多尺度区域合并分割方法和基于量化合并代价的快速区域合并分割方法；(4)针对矢量图形优化,提出了基于删除代价的矢量图形单层次优化方法、面向遥感影像矢量化图形的多层次优化方法和面向建筑群的矩形拟合优化方法；(5)针对三维信息提取,提出了基于扩展统计模型的建筑群高度提取方法和三种城市建筑群层数估算模型,针对侧向航拍影像提出了一种建筑群坐标修正方法；(6)提出了“参-建分离”的系统架构。针对该架构中的参数管理模块,设计了参数的关联、组织、管理和属性块的恢复、管理等方法。针对服务网站模块,设计了风格库管理和项目库管理子模块。针对自动建模模块,提出了DXF-SHP文件格式自动转换,CGA文法规则和规则库框架设计,规则库的调用和参数值传递,以及自动化建模脚本的设计等系列方法；(7)集成关键技术方法,开发了城市建筑群三维重建软件原型系统(3DRS)及其子系统CBRS、CityUp,并以杭州市西湖区为案例开展了实证研究,从精度、效率、成本、技术门槛、时效性等方面验证了整套解决方案的可行性。研究表明：(1)建筑物目标识别与参数化建模技术,恰可解决建筑物三维重建面临的两大难点；“三元交叉框架”使三大技术体系成为一个紧密连接、流程清晰、分工明确、目标一致的统一整体,为相关研究提供了理论支撑和方法指导；(2)面向对象的多尺度区域合并分割方法使综合考虑多种地物特征和多尺度分割成为可能,最大限度地缩短了队列长度,提高了分割精度和速度；基于量化合并代价的快速区域合并分割方法不仅能够保证分割精度,而且合并速度优势随初始分割区域数量的增加而越加显著；(3)相较于经典DP方法,基于删除代价的矢量图形单层次优化方法具有更高的精度和更低的时间复杂度,而且单位节点处理能力和等压缩率下的处理速度均具有显著优势；相较于传统单层次优化方法,面向遥感影像矢量化图形的多层次优化方法对影像不同的分割尺度和空间分辨率具有更强的适应性,能更好地还原地物的多层次特性；面向建筑群的矩形拟合优化方法可以有效减少最小面积外接矩形的计算时间,确保矩形对边平行、邻边垂直的关系,所得优化结果的形态和面积精度均较为理想；(4)基于扩展统计模型的建筑群高度提取方法、建筑群层数估算模型和建筑群坐标修正方法分别达到了较高的精度水平,可满足相关应用需求；(5)“参-建分离”的系统架构以及针对三大模块提出的一系列创新方法大幅降低了参数化建模平台的技术门槛和边际成本,提高了建模效率,为参数化技术的快速、广泛普及提供了新的发展思路；(6)本研究所提方法及系统在精度、效率、成本、门槛、时效性等方面均满足实验预定目标,体现了广泛的优势和大众化特性,整套解决方案具有可行性；(7)研究成果将推动数字城市三维空间数据基础设施的建设,使之在城市规划与管理、建筑景观设计、国防军事、应急救灾、环境保护、虚拟旅游、交通导航等诸多领域得到更广泛的应用。
[Abstract]:With the application of three-dimensional space of the fast development of digital city and city model of the continuous expansion of city buildings, the three-dimensional model of the growing demand. However, the traditional method of 3D building reconstruction in the face of large spatial scale, large amount of data, update the fast pace of city buildings, the efficiency, accuracy, cost, scale, technical barriers and so on are different degrees of defects. Therefore, this paper tries to introduce the target identification and parametric technology, to explore a suitable for city buildings with low cost, low threshold, high efficiency of the "popular" reconstruction solutions to meet the urgent needs of digital city and related fields.
The main contents of this paper include: (1) the building of 3D reconstruction, object recognition and related research of parametric modeling of the three fields, the technology method of the review, points out the existing problems; (2) of the three technology system for the depth of deconstruction, pointed out the causes of "two parallel framework" that has given the feasible way of the cross, and through the reconstruction of the system, the construction of city buildings reconstruction "three element cross framework"; (3) for remote sensing image segmentation, proposed multi-scale region merging object-oriented segmentation method and fast region merging cost based on quantitative combined segmentation; (4) according to the vector graphics optimization is put forward based on the vector graphics single level optimization method to remove the cost of remote sensing image vector graphics method of multi-level optimization and fitting optimization method for rectangular buildings; (5) according to the three letter Information extraction is proposed for estimating the model extended statistical model building height extraction methods and three kinds of city buildings based on the lateral aerial image layers, a modified method is proposed for building group coordinates; (6) proposed a system architecture to build parametric separation ". According to the parameters of the management module of the architecture, association the parameters of the design, organization, management and attribute block recovery, management and other methods. According to the service module, the design style of library management and project management module. In the automatic modeling module, the DXF-SHP automatic file format conversion, the design of library framework grammar rules and CGA rules, rules library calls and parameter values transfer and automatic modeling script design series method; (7) the key technology integration, the development of city buildings 3D reconstruction software prototype system (3DRS) and its sub system CBRS, CityUp, and Hangzhou The city Xihu District has carried out an empirical study for the case, verifying the feasibility of the whole solution from the aspects of accuracy, efficiency, cost, technical threshold and timeliness.
Research shows that: (1) modeling technology of building target recognition and parameter, it can solve the two difficulties of reconstructing buildings; "three element cross framework" the three technical systems become a closely connected, clear flow, clear division of labor, the overall goal consistent, provides theoretical support and guidance method for the relevant research; (2) multiscale region merging object-oriented segmentation method to integrate a variety of features and multi-scale segmentation becomes possible, minimizing the queue length, improve the accuracy and speed of segmentation; segmentation method combined with quantitative fast region merging based on cost can not only ensure the accuracy of segmentation, but also with the speed advantage with the increase in the number of initial segmentation region has become increasingly significant; (3) compared to the classical DP method has higher precision of vector graphics single level optimization method based on cost delete Degree and lower time complexity, and node processing capacity and processing speed of compression ratio has a significant advantage; compared with the traditional single level optimization method, segmentation scale and spatial resolution of multi-level optimization method for remote sensing image vector graphics on different images has stronger adaptability and better able to restore more than object oriented hierarchy characteristic; rectangle fitting optimization method can effectively reduce the calculation time of rectangle of minimum area outside the building, to ensure that the rectangular edge parallel relationship between adjacent vertical sides, the shape and area of accuracy of the optimization results are ideal; (4) extended statistical model extraction method based on height of buildings, building group number estimation model and complex coordinate correction method respectively, higher levels of precision, which can meet the application needs; (5) "Shen - built" separation system A series of innovative methods of system architecture and put forward three modules to greatly reduce the technical threshold of parametric modeling platform and marginal cost, improve the efficiency of modeling, parametric technology rapid popularization, provide a new development ideas; (6) of the proposed method and system in precision and efficiency. The cost threshold, timeliness meets the experimental target, embodies the advantages of widely and popular characteristics, set of feasible solutions; (7) study will promote the construction of digital city spatial data infrastructure, the city planning and management, landscape architecture, military defense, emergency relief, environmental protection, virtual tourism, can be widely used in many fields such as transportation and navigation.

【学位授予单位】：浙江大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：TU198;P208

【参考文献】