基于无人机激光雷达的不同种类红树林单木分割

红树林是生长在热带和亚热带海岸线上的典型潮间带植物,气候变化、自然灾害和人为干扰的影响下退化严重,加强红树林监测是有效保护红树林的重要工作。本文选取广东雷州市东海岸红树林示范区为研究区域,通过样地无人机激光雷达(LiDAR)高密度点云数据以及样方实测数据搭建红树林结构参数反演模型,提取不同红树林种类样方的单木结构参数并检验红树林点云分割精度,依据精度检验指标与偏差度量指标验证单木反演参数精度并分析原因。结果表明(1)基于分水岭算法对不同种类红树林实地样方进行单木树高提取,其中无瓣海桑的整体分割效果最好。单木树高参数精度分析与拟合验证,无瓣海桑结果最优,其次为秋茄,白骨壤精度最低。(2)通过估测单木冠幅反演值与样地实测值的偏差以及偏离比例,发现红树林单木冠幅参数整体反演效果较差,识别精度最高的还是无瓣海桑。(3)单木树高的识别效果整体优于单木冠幅,整体分割结果为无瓣海桑最高,其次为秋茄,白骨壤最低。使用无人机激光雷达技术可以有效提取红树林结构参数,替代传统的红树林人工测量方法,为红树林资源动态变化监测及保护提供高精度基础数据。

Study on single tree segmentation of mangrove trees of different species based on UAV LiDAR

Mangrove is a typical intertidal plant growing on tropical and subtropical coastlines. Mangrove is degraded seriously under the influence of climate change, natural disasters and human disturbance. Strengthening mangrove monitoring is an important work for effective protection of mangrove. In this paper, mangrove demonstration area in the East coast of Leizhou, Guangdong Province was selected as the research area. The inversion model of mangrove structural parameters was established by using high-density point cloud data of unmanned aerial laser radar (LiDAR) and measured quadrat data in the sample area, and the single wood structural parameters of different mangrove species quadrat were extracted and the segmentation accuracy of mangrove point cloud was tested. Verify the accuracy of single log inversion parameters and analyze the reasons according to the accuracy test index and deviation measurement index. The results showed that: (1) Based on the watershed algorithm, the height of individual tree was extracted from the field quadrat of different mangrove species, and the overall segmentation effect was the best. The accuracy analysis and fitting verification of high parameters of single tree showed that the results of A. japonicum were the best, followed by that of kandelia candel, and that of avicennia marina was the lowest. (2) By estimating the deviation and deviation proportion between the inversion value of single tree canopy width and the measured value in the sample site, it was found that the overall inversion effect of single tree canopy width parameter of mangrove was poor, and the highest recognition accuracy was found in Sonnerus apetala. (3) The recognition effect of single tree height was better than that of single tree crown width, and the overall segmentation result was the highest for sonneria apetala, followed by kandelia candel, and avicennia marina was the lowest. The UAV LiDAR technology can effectively extract mangrove structural parameters, replace the traditional mangrove manual measurement method, and provide high-precision basic data for the dynamic change monitoring and protection of mangrove resources.