Data from: Flying high: Sampling savanna vegetation with UAV-lidar
收藏资源简介：
The flexibility of UAV-lidar remote sensing offers a myriad of new opportunities for savanna ecology, enabling researchers to measure vegetation structure at a variety of temporal and spatial scales. However, this flexibility also increases the number of customizable variables, such as flight altitude, pattern, and sensor parameters, that, when adjusted, can impact data quality as well as the applicability of a dataset to a specific research interest. To better understand the impacts that UAV flight patterns and sensor parameters have on vegetation metrics, we compared 7 lidar point clouds collected with a Riegl VUX-1LR over a 300 x 300 m area in the Kruger National Park, South Africa. We varied the altitude (60 m above ground, 100 m, 180 m, and 300 m) and sampling pattern (slowing the flight speed, increasing the overlap between flightlines, and flying a crosshatch pattern), and compared a variety of vertical vegetation metrics related to height and fractional cover. Comparing vegetation metrics from acquisitions with different flight patterns and sensor parameters, we found that both flight altitude and pattern had significant impacts on derived structure metrics, with variation in altitude causing the largest impacts. Flying higher resulted in lower point cloud heights, leading to a consistent downward trend in percentile height metrics and fractional cover. The magnitude and direction of these trends also varied depending on the vegetation type sampled (trees, shrubs, or grasses), showing that the structure and composition of savanna vegetation can interact with the lidar signal and alter derived metrics. While there were statistically significant differences in metrics among acquisitions, the average differences were often on the order of a few centimeters or less, which shows great promise for future comparison studies. We discuss how these results apply in practice, explaining the potential trade-offs of flying at higher altitudes and alternating flight pattern. We highlight how flight and sensor parameters can be geared toward specific ecological applications and vegetation types, and we explore future opportunities for optimizing UAV-lidar sampling designs in savannas.
无人机激光雷达（UAV-lidar）遥感技术的灵活性为稀树草原生态学研究带来了诸多全新机遇，使研究人员能够在多样的时空尺度上开展植被结构的量化测量。但这种灵活性同时也提升了可自定义变量的数量，例如飞行高度、航线模式与传感器参数等；上述变量经调整后，既会影响数据质量，也会改变数据集对应特定研究方向的适用性。为更深入地明晰无人机飞行航线模式与传感器参数对植被指标的影响，本研究针对南非克鲁格国家公园内一块300×300米的研究区域，对比了7组采用Riegl VUX-1LR传感器采集的激光雷达点云数据。本研究中我们设置了多组飞行高度变量（离地60米、100米、180米与300米）与采样模式变量（降低飞行速度、增大航线间重叠度、采用交叉航线模式），并对比了多种与植被高度及植被覆盖占比相关的垂直植被指标。通过对比不同飞行模式与传感器参数下采集数据的植被指标，本研究发现飞行高度与航线模式均会对反演得到的植被结构指标产生显著影响，其中飞行高度变化带来的影响最为突出。飞行高度越高，点云反演得到的植被高度值越低，进而导致百分位高度指标与植被覆盖占比呈现出一致的下降趋势。这类趋势的幅度与方向还会因采样植被类型（乔木、灌木或草本）的不同而产生差异，这表明稀树草原植被的结构与组成会与激光雷达信号产生交互作用，进而改变反演得到的指标结果。尽管不同采集批次的植被指标间存在统计学意义上的显著差异，但平均差值通常仅为数厘米甚至更小，这为后续的跨批次对比研究提供了良好的应用前景。本研究探讨了上述结果的实际应用场景，阐释了提升飞行高度与切换航线模式可能带来的潜在权衡取舍。研究还阐明了如何将飞行与传感器参数适配于特定的生态学应用场景与植被类型，并探讨了优化稀树草原无人机激光雷达采样方案的未来发展方向。
