全波形激光雷达的波形优化分解算法

Optimization Decomposition Method of Full-waveform LiDAR

随着数据存储能力和处理速度的提高,三维激光扫描系统逐渐具备全波形采集和分析技术。为了从全波形数据中获得脉冲时间、幅度、脉宽以及多回波分布等综合信息,波形分解成为了全波形激光雷达数据处理的关键技术之一。针对LM算法在一定程度上依赖初值,而传统激光雷达数据处理容易遗漏部分重叠的返回波,本文提出了一种改进回波分量初值设定的算法来获取回波脉冲的位置、宽度和强度。针对一套自主研发的全波形记录激光雷达演示系统进行了波形分解试验,定性和定量分析结果验证了该方法的有效性、可靠性和准确性。

With the improvement of data storage capacity and data processing capabilities,full waveform LiDAR develops rapidly and then from waveform data abundant information about the physical characteristics of the targets can be effectively retrieved through data processing.Waveform decomposition is therefore the key to full waveform LiDAR data processing.However,the number and initial parameters of the echo components are difficult to set in waveform decomposition.Conventional decomposition methods detect echo components by peak points or using the threshold method,which may ignore some overlapping components and show low accuracy.To this end,in this paper a novel method based on LM algorithm that takes into account both peak points and inflection points is adopted and it can extract the location,amplitude and FHWM of the echo components,proving it is a reliable and high accurate decomposition algorithm.To further demonstrate the advantages of the suggested method,waveform data measured by a full waveform LiDAR demonstration system and generated from simulation were both decomposed using the method.The results show that the suggested algorithm is efficient,promising and can effectively decompose adjacent echo components,then will improve the accuracy in the next phase of data processing.