计算机工程与应用 ›› 2012, Vol. 48 ›› Issue (33): 183-187.

• 图形、图像、模式识别 • 上一篇    下一篇

基于第二代Curvelet变换的点云曲面特征提取

杨红娟1,陈继文2,张运楚1   

  1. 1.山东建筑大学 信息与电气工程学院 山东省智能建筑技术重点实验室,济南 250101
    2.山东建筑大学 机电工程学院,济南 250101
  • 出版日期:2012-11-21 发布日期:2012-11-20

Surface feature extraction based on curvelet transform from point cloud

YANG Hongjuan1, CHEN Jiwen2, ZHANG Yunchu1   

  1. 1.Shandong Provincial Key Lab of Intelligent Buildings Technology, School of Information & Electrical Engineering, Shandong Jianzhu University, Jinan 250101, China
    2.School of Mechanical and Electrical Engineering, Shandong Jianzhu University, Jinan 250101, China
  • Online:2012-11-21 Published:2012-11-20

摘要: 针对点云特征提前取方法在多方向性分析方面的局限性,将Curvelet变换引入点云的分析,研究数据点云不同尺度曲面特征的提取方法。在数据点云分层、扩展预处理的基础上,以第二代离散Curvelet变换分析数据点云,采用软硬阈值折衷法,对表示数据点云边缘的Detail层、Fine层Curvelet变换系数进行处理,增强数据点云的边缘。对增强后的Curvelet变换系数进行Curvelet逆变换,重构数据点云,提取数据点云的边缘,获取曲面特征。实例表明,以Curvelet变换分析为基础的曲面特征提取方法,可以更加准确地提取数据点云的曲面特征。

关键词: 数据点云, 特征提取, 多尺度几何分析, Curvelet变换

Abstract: The current surface feature extraction method has limitations for multi directions analysis of data point. Curvelet transform is introduced to analysis of data point to research on surface feature extraction method. Second-generation discrete curvelet transform is used to analyze data point based on slicing and expanding preprocessing data point. Curvelet transform coefficients of detail layer and fine layer present the edge and contour of data point. Compromise for soft and hard thresholds is used to process curvelet transform coefficients of Detail layer and Fine layer to enhannce the edge and contour of data point. The data point is reconstructed from the enhanced Curvelet transform coefficient with Curvelet inverse transformation. Boundary contour is achieved to extract surface feature. Surface feature is achieved by extrcating the edge and contour of data point. A surface feature extraction method is proposed based on Curvelet transform. Example shows the proposed method can accurately extract surface feature from data point.

Key words: data point, feature extraction, multi-scale geometric analysis, Curvelet transform