房屋建筑图形数据多通道整合方法设计与仿真  

Design and Simulation of Multi-Channel Integration Method for Building Graphic Data

作　　者：高峻峰 汪新军[2] GAO Jun-feng;WANG Xin-jun(Huazhong University of Science and Technology,School of Mechanical Science&Engineering,Wuhan Hubei 430074,China;Huazhong University of Science and Technology,Wuhan Hubei 430074,China)

机构地区：[1]华中科技大学机械科学与工程学院,湖北武汉430074 [2]华中科技大学,湖北武汉430074

出　　处：《计算机仿真》2022年第6期496-500,共5页Computer Simulation

摘　　要：海量数据属性复杂、影响整合效果,传统数据整合方法缺少对图形数据的预处理,不完整的编码影响数据整合与共享效果。研究统一编码下房屋建筑图形数据多通道整合方法。预处理房屋建筑图形数据,补充与更新统一编码,构建模糊集合挖掘房屋建筑图形数据特征,通过构建图形特征的多通道卷积神经网络模型,实现对图形数据的多通道交互控制与整合。实验结果表明,与传统方法相比,文中方法生成的空间数据整合效果图更完整,传统方法生成的效果图存在部分缺失。人为完善统一编码后,评价不同方法整合数据过程中共享数据的能力,发现文中方法的数据共享能力评价结果优于传统方法,可见所提方法更具备使用意义。The complex mass data attributes affect the integration effect of graphics, and the lack of graphics data preprocessing leads to the poor integration effect of traditional data integration methods. Therefore, this paper studies the multi-channel integration method of building graphic data under unified coding. The building graphic data was preprocessed. Fuzzy set mining house building graphic data features were established via supplementing and updating unified coding. The multi-channel convolution neural network model of graphic features was built to multi-channel interactive control and integrate graphic data. The experimental results show that the evaluation results of the integrity and data sharing ability of the integration effect map of this method are better than the traditional methods.

关 键 词：统一编码 房屋建筑 图形数据 多通道整合 多通道卷积神经网络 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]