论文部分内容阅读
本文提出了一种估计多个三维刚体运动参数的鲁棒（ｒｏｂｕｓｔ）算法，可以处理包含高斯噪声和出格点（ｏｕｔｌｉｅｒ）的对应点数据．根据贝叶斯统计决策规则和蕴含在问题中的启发式规则，我们将运动参数估计问题转化为极大似然估计过程，实现部分模型拟合（ｐａｒｔｉａｌｍｏｄｅｌｆｉｔｔｉｎｇ）．因此，这种优化算法就是估计一组三维运动参数，使对应点数据最大限度地拟合似然函数，从而保证算法的鲁棒性．
In this paper, we propose a robust algorithm to estimate the motion parameters of multiple 3D rigid bodies, which can process the corresponding point data containing Gaussian noise and outlier. According to the Bayesian statistical decision rules and the heuristic rules implicated in the problem, we transform the motion parameter estimation problem into the maximum likelihood estimation process and realize the partial model fitting. Therefore, this optimization algorithm is to estimate a set of three-dimensional motion parameters, so that the corresponding point data to maximize the likelihood function, thus ensuring the robustness of the algorithm.