点云建模与空气动力学的橡胶林台风损伤机制与安全分析

Wind is one of the most detrimental disturbance agent for rubber plantations in the China’s southeast coastal areas. In this paper, we focus on the vulnerability assessment and safety evaluation for rubber trees under wind disturbance impacts through employing laser scanning technique and aerodynamics theories. First, laser scanning technique and physical measurements are taken as the primary means, combined with computer vision and graphics algorithms to automatically derive biological growth properties and structural parameters of various organs of rubber trees. Second, using aerodynamics theories to simulate turbulence wind field and modeling wind loads on rubber trees with various species and different planting structure are the main purpose to investigate wind hazards on plants in order to analyze the biomechanics of wind-plant interaction and account for the relationship between damage rates caused by wind and stand variations in e.g. tree properties and spatial distribution. Third, combination of manual inventory, laser scanning and multi-source remote sensing are adopted to monitor spatio-temporal dynamics of wind disturbance impacts on rubber plantations and quantitatively obtain ecological system description and feedback of rubber trees at different scales. The assessment of typhoon damage on forest derived from various measurement data at different resolution also can be verified and coupled with each other. At last, effective measurement devices for the wind-related data acquisition are used for our experiment validation. Meanwhile, man-made small tree 3D models are adopted to compose various rubber plantation models which are taken as the field data putting into the wind tunnel experiment instrument for conducting validation experiments in order to assess if the results presented in our project are valid at full scale. Our experimental site located mainly in Danzhou city, which is largest rubber production base on Hainan Island, China. This project explore hurricane protection and wind-related disaster risk reduction strategies for the rubber trees in China’s southeast costal region, and the result will provide reliable decision-making basis for scientific management and practical guidance of rubber plantation planting.

本课题面向中国沿海橡胶林木受风害问题，拟建立基于激光点云与空气动力学的橡胶林木风害评估与安全评价体系。首先，采用激光雷达与物理测量作为橡胶林木参数的估测手段，并融合视觉计算与图形学方法构建具有真实器官参数与几何形态的橡胶林木的机体模型。其次，基于空气动力学理论构造热带强气流气旋下空间风场模型，并找寻适用于不同品系及种植结构下橡胶林风力阻遏下的流固耦合模型与产生的生物动力学行为，探明不同品系及林学参量与风力侵害所造成的损伤级别因果关系。再次，通过林调、激光点云、多源遥感数据联合分析，在不同尺度中对中国沿海橡胶林地域内的时空推演与生态体系变迁进行定量定性描述，进而研究不同尺度间台风灾害损伤评估的契合与互映射关系。最后，运用真实风场测绘仪器与风洞实验对林分在风力载荷下进行实景测量与结果验证。本课题以海南儋州橡胶林场为主试验区，服务于中国沿海橡胶林的防灾减灾问题，具有重要的理论支撑与现实指导意义。

本课题面向中国沿海橡胶林木受风害问题，建立基于激光点云与空气动力学的橡胶林木风害评估与安全评价体系。首先，采用激光雷达与物理测量作为橡胶林木参数的估测手段，并构建了基于人工智能的深度学习网络与基于计算机图形学的视觉计算方法，针对采集的橡胶林激光点云数据开展枝叶分离、林木骨架提取、叶团簇大小计算、透光率反演、叶叶分离与单叶形态阐释等精细参数的提取。其次，提出了基于空气动力学理论构造热带强气流气旋下空间风场模型，并应用于不同品系及种植结构下橡胶林中，分析风与林木之间的流固耦合模型，来计算林场内的风压、风速、湍流强度等关键参量，为林木的防灾减灾提供科学指导。再次，通过真实林调数据，并联合Landsat 7/8和Sentinel-2时间序列遥感影像和影像差异法评估橡胶林风害损失量，通过分析2019年8月1912号热带风暴杨柳在海南西部诱发的龙卷风造成海南西部约645公顷的橡胶林倒伏，说明我们联合Landsat和Sentinel-2监测橡胶林风害损失的可行性。最后，搭建了真实一千公斤的大型试验风洞设备，并结合其他风速测量传感器对橡胶林分在风力载荷下进行实景测量、计算并与提出的方法结果对比验证，同时，真实的获取不同风级下不同位置的橡胶林木的冠形与冠积变化、树体倾斜度、叶面积衰减等关键参量。本课题以海南儋州橡胶林场为主试验区，服务于中国沿海橡胶林的防灾减灾问题，具有重要的理论支撑与现实指导意义。