基于组合光透视的室内三维空间分割方法  

Indoor Three-dimensional Space Segmentation Method Based on Combined Light Perspective

作　　者：赵亚琴[1] 蔡晓骝[1] ZHAO Ya-qin;CAI Xiao-liu(School of Art and Design,Anhui Technical College of Mechanical and Electrical Engineering,Wuhu 241000,China)

机构地区：[1]安徽机电职业技术学院艺术设计学院,安徽芜湖241000

出　　处：《辽东学院学报（自然科学版）》2022年第1期42-47,共6页Journal of Eastern Liaoning University:Natural Science Edition

基　　金：安徽省教育厅人文社科重点项目(SK2019A0978)。

摘　　要：室内三维空间分割过程中存在障碍物遮挡时,传统分割方法容易陷入局部循环迭代导致空间建模出现偏差,为此,设计了一种基于组合光透视的室内三维空间分割方法。首先,在组合光透视技术下提取室内空间点云,布设采集点并对得到的点云进行去噪;其次,利用得到的点云数据进行室内三维空间建模,优化像素点,利用语义分割结果修复深度图像;最后,利用射线追踪优化室内三维空间分割算法,设计空间分割方法流程。为验证设计方法的可靠性,随机选择某建筑中的室内空间进行分割方法验证,实验结果表明,利用设计的方法获取的点云数据和实际数据之间误差低于1%,并且得到的特征点与实际的室内特征点的误差距离更小,从而验证了方法的有效性。When there are obstacles in the process of indoor three-dimensional space segmentation,the traditional segmentation method is easy to fall into local loop iteration,which leads to spatial modeling deviation.Therefore,an indoor three-dimensional space segmentation method based on combined perspective is designed.The indoor space point cloud is extracted under the combined perspective technology,the acquisition points are laid and the obtained point cloud is denoised.The obtained point cloud data are used to model the indoor three-dimensional space,optimize the pixels,repair the depth image by semantic segmentation results,optimize the indoor three-dimensional space segmentation algorithm by ray tracing,and design the space segmentation method flow.In order to verify the reliability of the design method,the indoor space in a building is randomly selected to verify the segmentation method.The experimental results show that the error between the point cloud data obtained by the design method and the actual data is less than 1%,and the error distance between the obtained feature points and the actual indoor feature points is smaller,which verifies the effectiveness of the method.

关 键 词：组合光透视 室内三维空间 点云提取 空间分割 特征点匹配 

分 类 号：TN929[电子电信—通信与信息系统]