aIB算法在古建筑信息模型特征提取中的应用与研究  

Application and research of aIB algorithm in ancient building information model feature extraction

作　　者：宋阳[1,2] 李昌华[1] 马宗方[1] 李智杰[1,2] SONG Yang;LI Changhua;MA Zongfang;LI Zhijie(College of Information and Control Engineering, Xi'an Univ. of Arch. & Tech. Xi'an 710055, China;College of Architecture, Xi'an Univ. of Arch.& Tech., Xi'an 710055, China)

机构地区：[1]西安建筑科技大学信息与控制工程学院,陕西西安710055 [2]西安建筑科技大学建筑学院,陕西西安710055

出　　处：《西安建筑科技大学学报（自然科学版）》2016年第4期606-609,616,共5页Journal of Xi'an University of Architecture & Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61373112,50878176);西安建筑科技大学人才科技基金项目(RC1343)

摘　　要：古建筑信息模型特征提取在古建筑重建过程中具有重要的作用.针对三维激光扫描获取的散乱点云数据形成的古建筑信息模型在特征提取时存在的问题,采用二值化方法将图像数字化,提取特征数据集,使用aIB算法提取点云数据特征.aIB算法将原始点云数据集作为源变量X,将法向量作为相关变量Y,将特征变量作为目标变量T,将源变量X压缩到目标变量T中时,尽可能保持相关变量Y的信息,更加精确的获取点云数据特征,同时尽量压缩噪声数据.实验结果表明aIB算法可以有效提升古建筑信息模型特征提取的准确度.In the reconstruction ofancient building, The information model of feature extraction of the building plays an importantrole. By 3D laser scanning to obtain scattered point cloud data of the ancient building information model, aiming at the problem ofdata in the feature extraction, the image binarization method is used to extract features of data set, using aIB algorithm to extract thecharacteristics of point cloud data. In aIB algorithm, the original point cloud data is set as the source variable X; the normal vector isset as the relevant variables Y and characteristic variables as target T. The source variable X is compressed to the target variable T,and at the same time, relevant variables Y should maintain accurate information as possible, so as to accurately obtain point clouddata characteristics and also try to compress the data noise. The experimental results show that aIB algorithm can effectively improvethe accuracy of the feature extraction in ancient building information model.

关 键 词：古建筑信息模型 特征提取 互信息 AIB 

分 类 号：TU18[建筑科学—建筑理论]