《基于图形理解的平法图钢筋信息自动获取方法研究》由会员分享，可在线阅读，更多相关《基于图形理解的平法图钢筋信息自动获取方法研究（67页珍藏版）》请在金锄头文库上搜索。
1、太原理工大学 硕士学位论文 基于图形理解的平法图钢筋信息自动获取方法研究 姓名：李华 申请学位级别：硕士 专业： 指导教师：陈俊杰 20100323 太原理工大学硕士研究生学位论文 I 基于图形理解的平法图钢筋信息自动获取方法研究 摘 要 基于图形理解的平法图钢筋信息自动获取方法研究 摘 要 我国为与国际先进的设计方法接轨，推广实施了新的制图标准“混凝 土结构施工图平面整体表示方法”（简称“平法”） 。该方法在设计上具有 很多优点但是由于设计的简化，带来了施工、监理和工程预算工作的复杂 化，因而,计算机辅助建筑算量技术的研究及相关软件的开发受到了越来越 多的重视。 本论文在分析了当前计算机辅助
2、建筑算量技术的研究现状及其优缺点 的基础上，对钢筋信息的获取方法展开研究。提出了钢筋信息自动获取的 方法和建立用于存储钢筋信息的钢筋表达模型的方法，并用实验对其可行 性予以证明。 论文的主要工作有以下几点： 1、基于图形理解及平法图制图规则，对建筑图进行语义分析。 平法图主要由图形元素、尺寸标注和文字说明三部分组成。这三部分 按照一定组合所共同表达的信息构成了平法图的主要内容。要想从静态的 图中提取出含有工程意义的信息首先就要对平法图进行语义分析，本文从 平法图的表达方式、投影关系和尺寸约束三个层次上进行了语义分析，得 出平法图的视图特点，这是研究自动获取钢筋信息方法的基础。 2、在语义分析的
3、基础上提出从平法图中自动获取钢筋信息的方法及相 太原理工大学硕士研究生学位论文 II 关算法： 提出图形特征类的概念，通过图形特征进行图形元素分离，作为钢 筋信息获取的切入点； 通过关联视图确定多张图纸的组合方式，得到完整的图形元素集； 通过分析标注文本和图形元素之间相对应的关系，构建尺寸关系树， 得到钢筋的尺寸信息。 3、建立了用于存储钢筋信息的钢筋信息表达模型。 本文对钢筋信息的注写规则进行详细的分析和研究，完整的归纳出钢 筋表达信息的特点，在此基础上建立了柱钢筋信息的表达模型，把描述柱 钢筋的信息按照所归纳出的表达特点分别封装在相应的类中，便于信息的 存储和处理。 4、通过实验，证明本文
4、提出的自动获取钢筋信息方法可以有效的完成 钢筋信息的提取任务。 关键词：图形理解，信息自动获取，平法图语义分析，信息表达模型 太原理工大学硕士研究生学位论文 III Research on Automatic Acquisition of Reinforcement Information Plane Whole Show Method Information Based on Graph Understanding ABSTRACT In order to get in line with international advanced design method, our country p
5、romotes the implementation of “Structure of concrete construction drawing plane whole show method”. This method have many advantages in the design, however, due to its design simplification, it bring complication on construction, supervisor and construction budget work. Therefore, research on Comput
6、er-aided quantity technology of contruction and development of related sofeware is attracting more and more attention. This thesis research on reinforcement information acquisition method based on analysis of study situation, advantages and disadvantages of Computer-aided quantity technology of cont
7、ruction. Besides, this thesis put forward Automatic Acquisition of Reinforcement Information Strategy and construction of reinforcement information expression models. The components of this thesis are as follows 1. Based on image understanding and semantic analysis of Drawing Plane Whole Show Method
8、 Graphical Law, author put forward Automatic Acquisition of Reinforcement Information Stragety. 太原理工大学硕士研究生学位论文 IV Plane Whole Show Method Graph include graphic elements, dimension and textual description. According to certain combination, the information that three parts convey constitute main cont
9、ent of Plane Whole Show Method Graph. If want to extract significant engineering information, we must do semantic analysis to Plane Whole Show Method Graph at first. This thesis conducted semantic analysis from expression, projection relation and dimension restriction of Plane Whole Show Method Grap
10、h to reach the view characteristic, then author put forward Automatic Acquisition of Reinforcement Information Stragety. 2. Method to acquire reinforcement information by Graphics Feature Understanding the Plane Whole Show Method Graph should be guided by certain engineering knowledge. This research
11、 separated construction renforcement graph data through analyzing graphics feature and putting forward the concept of gragh characteristic classes, make which to be reinforcement information acquisition pointcut. 3. Method to acquire reinforcement information by view context Shape information which
12、reflect functional unit of same construction is not in same view. This research analyzed view presentation of Plane Whole Show Method Graph and understood how so many graph paper combined with each other to express reinforcement information completely. Based on analysis of interrelationship between
13、views, author put forward method to acquire reinforcement information by view context and specific algorithm. 太原理工大学硕士研究生学位论文 V 4. Method to acquire reinforcement dimension information by labeling functions Labeling functions play an important part in dimension information extraction. Semantic under
14、standing of Labeling functions is very important in Plane Whole Show Method Graph recognition. This research ocnstructed dimension relation tree. 5. Construction of reinforcement information expression model This research analyzed and studied label rules of reinforcement information and concluded re
15、inforcement presenting information characteristic. Then, author constructed presenting information model of column reinforcement, encapsulated presenting information of column reinforcement in the corresponding categories according to conclusion of presenting information characteristics, and stored
16、the information according to column reinforcement Dimension constraint, parametric representation of label information and construction information for convenience of processing information. KEY WORDS: image understanding, Information Extraction, semantic analysis of Drawing Plane Whole Show Method Graph, information expression model 太原理工大学硕士研究生学位论文 1 第一章 绪论 在建筑业的计算机应用中，CAD 技术的发展比较快，水平
