论文部分内容阅读
提出一种激光点云数据关联决策算法.基于判别图模型,提取并智能管理激光点云的多重形状特征,通过最大伪似然学习优化局部特征和配对特征的权重;应用最大和概率推理实现对图模型隐节点状态的估计,进而将激光点关联映射为最大后验概率的配置回溯问题;实验结果验证了所提出算法比传统算法具有更好的性能.
A decision-making algorithm based on laser point cloud data was proposed. Based on the discriminant graph model, the multi-shape features of the laser point cloud were extracted and intelligently managed, and the weights of local features and matching features were optimized by the maximum pseudo-likelihood learning. The model is used to estimate the state of hidden nodes and then the mapping of laser points is mapped to the configuration backtracking problem of maximum a posteriori probability. Experimental results show that the proposed algorithm has better performance than the traditional algorithm.