论文部分内容阅读
双通道数据拼接技术能有效地增大激光雷达系统的动态探测范围,改善大气回波信号的信噪比。提出了一种基于非支配排序遗传算法Ⅱ(NSGA-Ⅱ)与邻域粗糙集(NRS)的大气遥感激光雷达数据拼接智能算法,即NRSWNSGA-Ⅱ,提高了数据拼接的准确性和稳定性。该算法以三个判断双通道数据拟合优度的评价函数为优化目标,通过NSGA-Ⅱ获得评价函数的Pareto最优解集,进而利用NRS训练数据样本得到的权重进行线性规划,实现了最优拟合范围内的全局随机搜索。实验结果表明,所提算法拼接效果良好并在全天性数据拼接工作中有较好的稳定性。
Dual-channel data splicing technology can effectively increase the dynamic detection range of the lidar system and improve the signal-to-noise ratio of the atmospheric echo signal. This paper presents a novel remote sensing lidar data splicing algorithm named NRSWNSGA-Ⅱ based on non-dominated ranking genetic algorithm Ⅱ (NSGA-Ⅱ) and neighborhood rough set (NRS), which improves the accuracy and stability of data splicing. The algorithm uses three evaluation functions that determine the goodness of fit of two-channel data as the optimization objective, obtains the Pareto optimal solution set of the evaluation function through NSGA-Ⅱ, and then makes use of the weights obtained from the NRS training data samples for linear programming, Fitting within the scope of the global random search. Experimental results show that the proposed algorithm has good splicing effect and good stability in all-day data splicing.