基于改进微分增强法的激光雷达云参数反演算法

微分法是激光雷达云参数反演的一类传统方法，本文在微分增强法的基础上，优化一阶与二阶微分信号的拟合点数，平衡失真与噪声对检测结果的影响。以距离修正的回波信号代替云峰与云边界函数中的原始回波信号，增强云区域与非云区域的对比。对于因低层云存在而造成二次阈值过大的问题，通过扩大一次阈值的排除区间，有效减少对云层的漏判情况。采用激光雷达数据进行的实验结果表明，与改进前的算法相比，云底高度的相关系数由0.8930提高至0.9328,均方根误差由0.4924 km降低至0.2991 km,云顶高度的相关系数由0.8174提高至0.8598,均方根误差由0.7637 km降低至0.5912 km,改进后的算法具备更佳的反演结果。

LiDAR cloud parameter inversion algorithm based on improved differential enhancement method

The differential method is a traditional method for LiDAR cloud parameter inversion.In this paper,based on the differential enhancement method,the numbers of fitting points of the first order and second order differential signals are optimized to balance the effects of distortion and noise on the detection results.The distance corrected echo signal is used to replace the original echo signal in the cloud peak and cloud boundary functions to enhance the contrast between the cloud and non cloud regions.For the problem of excessive secondary thresholds due to the presence of low level clouds,the exclusion interval for primary thresholds is expanded to effectively reduce the missed detection of clouds.The experiment results using LiDAR data show that the correlation coefficient of cloud bottom height is improved from 0.8930 to 0.9328,and the root mean square error is reduced from 0.4924km to 0.2991km.The correlation coefficient of cloud top height is improved from 0.8174 to 0.8598,and the root mean square error is reduced from 0.7637km to 0.5912km,which proves that the improved algorithm has better inversion results.