| 星载激光测高仪多模式回波参数提取方法（特邀） |
| |
| 引用本文： | 朱天豪,周辉,石岩,张千胤.星载激光测高仪多模式回波参数提取方法（特邀）[J].红外与激光工程,2022,51(1):20210836-1-20210836-11. |
| |
| 作者姓名： | 朱天豪 周辉 石岩 张千胤 |
| |
| 作者单位： | 武汉大学 电子信息学院，湖北 武汉 430072 |
| |
| 基金项目： | 国家自然科学基金(41971302)；国防科工局十三五民用航天技术预先研究项目(250000887) |
| |
| 摘 要： | 全波形星载激光测高仪的接收波形特征参数可以用于反演目标的形貌信息，传统的波形处理算法不能用于混叠严重以及偏离高斯形态的多模式波形特征参数提取。针对混叠严重的多模式回波，提出一种基于偏正态拟合模型，使用激励Richardson-Lucy反卷积算法、逐层分解算法、梯度下降法和非线性最小二乘拟合算法相组合的波形特征参数提取方法。采用已知参数的波形数据集、机载仿真波形数据集和全球生态系统动态调查（GEDI）激光雷达波形数据，基于波形相关系数与均方根误差（RMSE）、波形特征参数相对误差、波形分量个数提取正确率等评价指标开展波形处理试验，并将处理结果与传统的高斯分解结果进行比较分析。已知参数波形数据集处理结果的平均波形相关系数提升了约2%，RMSE降低了约47%，波形特征参数相对误差平均降低了约5%，分量个数提取正确率提升了约34%；机载仿真数据和GEDI波形数据处理结果的平均波形相关系数分别提升了约1%和2%，RMSE分别降低了约56%和54%。同时，开展了陡坡区域植被高度解算的仿真试验，得到的植被高度准确程度明显高于传统方法。所有处理结果均表明该方法更有利于多模式回波特征参数的提取以及目标参数的反演。
|
| 关 键 词： | 全波形激光测高 多模式回波 特征参数提取 偏正态模型 激励反卷积 |
| 收稿时间： | 2021-11-10 |
Parameter extraction method on the multiple mode waveforms of satellite laser altimeter(Invited) |
| |
| Zhu Tianhao,Zhou Hui,Shi Yan,Zhang Qianyin.Parameter extraction method on the multiple mode waveforms of satellite laser altimeter(Invited)[J].Infrared and Laser Engineering,2022,51(1):20210836-1-20210836-11. |
| |
| Authors: | Zhu Tianhao Zhou Hui Shi Yan Zhang Qianyin |
| |
| Affiliation: | School of Electronic Information, Wuhan University, Wuhan 430072, China |
| |
| Abstract: | The parameters of the received waveform of a full-waveform satellite laser altimeter can be used for retrieving the morphological information of the target. The traditional waveform processing algorithm is unable to extract the parameters for a non-Gaussian and overlapped multiple-mode waveform. Therefore, a synthetic algorithm with the boosted Richardson-Lucy deconvolution, layered extraction, gradient descent and nonlinear least square was proposed for a skew-normal full-waveform decomposition. The proposed waveform processing experiments were implemented by employing the known-parameter waveforms, airborne simulated waveforms and global ecosystem dynamics investigation (GEDI) lidar waveforms and the evaluation indictors including the waveform correlation coefficient, root mean square error (RMSE), relative error of characteristic parameters, successful detection rate of the number of components. The processed results were compared with those by the traditional Gaussian decomposition algorithm. The average correlation coefficient of the processing results for the known-parameter data set had a growth of 2% and the average RMSE has a reduction of 47%. The average relative error of parameters was reduced by about 5% and successful detection rate of the number of components was improved by about 34%. For the simulated and GEDI lidar waveforms, the average correlation coefficients had slight growth of 1% and 2%, and the average RMSEs had a more significant reduction of 56% and 54%, respectively. In addition, the simulated verification of canopy height in steeped region was carried out. The precision of the derived canopy height was significantly higher than that of the traditional method. All processed results demonstrate that the proposed method is more conducive to the extraction of the multiple-mode waveform parameters and the inversion of target parameters. |
| |
| Keywords: | |
| 本文献已被 万方数据 等数据库收录！ |
| 点击此处可从《红外与激光工程》浏览原始摘要信息 |
|
点击此处可从《红外与激光工程》下载全文 |
|
