改进ICP算法在桥梁三维激光扫描数据配准中的应用研究

Application Research of Improved ICP Algorithm on Bridge 3D Laser Scanning Data Registration

为了解决传统ICP算法在桥梁三维激光扫描数据配准中迭代次数多、收敛速度慢的问题,基于局部拟合平面投影法搜索最近点以及幂法解算单位四元数,提出了一种优化配准速度与精度的改进ICP配准算法。在Microsoft Visual Studio2012平台运用C++编程语言将所提算法程序化,利用Lecia HDS8400三维激光扫描仪对某桥梁进行实验,通过比较传统ICP算法与改进ICP算法的配准效果及计算其配准残差可知,改进的ICP配准算法的套合效果优于传统ICP配准算法,收敛速度快、迭代次数少。研究成果可为获取点云在数据配准中提供有力技术借鉴。

The traditional ICP algorithm in the registration of bridge 3D laser scanning data has the problems of large number of iterations and slow convergence speed.Based on the local fitting plane projection method for searching the nearest point and the power method for solving the unit quaternion,an optimized ICP registration algorithm with improved speed and accuracy is proposed.The algorithm was developed using C++programming language on the Microsoft Visual Studio 2012 platform,and the Lecia HDS8400 threedimensional laser scanner was used to collect the laser points on a bridge.By comparing the registration residuals between the traditional ICP algorithm and the improved ICP algorithm,the improved ICP registration algorithm is better than the traditional ICP registration algorithm,with fast convergence speed and fewer iterations.The research results can provide a powerful technical reference for point cloud registration.