基于并行粒子群优化的三维点云配准算法

针对基于群智能优化的点云配准算法计算时间长的问题,提出一种基于CUDA的并行粒子群配准算法.以点对点距离最短为适应度函数,利用粒子群算法各粒子天然的并行性,将运算过程分配到GPU的各个线程中计算变换参数.由于GPU多个线程运算同时执行互不干扰,极大地提高了粒子群的运算速度,从而可以实现点云的快速、精确配准.实验结果表明,该算法既克服了ICP算法对点云初始位置要求高的缺点,又有效解决了基于群智能优化的点云配准算法计算时间长的问题.

Research of 3D Point cloud data registration based on parallel particle swarm optimization algorithm

To surmount the limitation of long computing time of point cloud registration based on swarm intelligence optimization algorithm,this paper proposes a parallel particle swarm optimization algorithm based on CUDA.Regarding the shortest distance between point and point as the fitness function, uterlizing the parralism of particle swarm optimization algorithm,operation process is distributed to various threads of GPU and calculate the transformation parameters, to realize the precise registration of point cloud. As implementation of multiple threads operation at the same time do not interfere with each other, which greatly improves the operation speed of particle swarm optimization. The experimental results show that the algorithm not only overcomes the disadvantage of the ICP algorithm with high requirement to the initial position of point cloud, but also effectively solves the problem of operation swarm intelligent algorithm which costs too much time.