长江科学院院报 ›› 2016, Vol. 33 ›› Issue (11): 1-4.DOI: 10.11988/ckyyb.20160811

• 遥感技术应用 • 上一篇    下一篇

双阈值Alpha Shapes算法提取点云建筑物轮廓研究

李云帆1,谭德宝1,高 广2,刘 瑞3,4   

  1. 1.长江科学院 空间信息技术应用研究所,武汉 430010;
    2.深圳飞马机器人科技有限公司,广东 深圳 518000;
    3.哈尔滨工业大学 深圳研究生院,广东 深圳 518055;4.深圳市房地产评估发展中心,广东 深圳 518040
  • 收稿日期:2016-08-10 出版日期:2016-11-20 发布日期:2016-11-08
  • 作者简介:李云帆(1984-),男,湖北恩施人,工程师,博士,主要从事机载、车载激光雷达点云数据处理方向研究,(电话)13429843035(电子信箱)yun_di@sina.com。
  • 基金资助:

    中央级公益性科研院所基本科研业务费项目(CKSF2014031/KJ);云南省水利重大科技项目(CKSK2015852/KJ)

Extraction of Building Contour from Point Clouds Using DualThreshold Alpha Shapes Algorithm

LI Yun-fan1,TAN De-bao1,GAO Guang2,LIU Rui3,4   

  1. 1.Spatial Information Technology Application Department, Yangtze River Scientific Research Institute, Wuhan 430010, China;
    2.Shenzhen Feima Robotics Co., Ltd., Shenzhen 518000, China;3.Shenzhen Graduate School, Harbin Institute of Technology,Shenzhen 518055, China;4.Center for Assessment and Development of Real Estate Shenzhen, Shenzhen 518040, China)
  • Received:2016-08-10 Online:2016-11-20 Published:2016-11-08

摘要:

针对单一阈值的Alpha Shapes算法在提取点云建筑物轮廓时存在的轮廓精度和完整性难以兼顾的问题,提出一种双阈值的Alpha Shapes算法,利用简单环的概念设计轮廓搜索算法,获得既有较好完整性又有较高几何精度的建筑物轮廓线;然后,利用一种最小二乘的轮廓线化简算法对提取出的初始轮廓进行化简,与经典的Douglas Peucker算法相比,在存在噪声的情况下,该方法化简后的轮廓线更接近实际的轮廓线。

关键词: LiDAR, 建筑物轮廓提取, RANSAC, Alpha Shapes算法

Abstract:

To balance the contour accuracy and completeness of single threshold Alpha Shapes in extracting point cloud building contours, we present a dual-threshold Alpha Shapes algorithm using a simple ring design concept contour search algorithm to obtain both a good integrity and a relatively high geometric precision of the building’s contour. Furthermore, the initial contour is simplified based on least squares algorithm. In the presence of noise, the simplified contour lines of the present algorithm are closer to the actual contours compared with the classic Douglas Peucker algorithm.

Key words: LiDAR, building boundaries extraction, RANSAC, Alpha Shapes algorithm

中图分类号: