Tree height is one of the most important tree attributes in forest inventory. However, using conventional field methods to measure tree height is a laborious and time-consuming process. Despite the great interest in the past to facilitate tree height measurements, new, upcoming solutions are not yet thoroughly investigated. In this study, we investigated the applicability of different close-range remote sensing options for tree height measurement in a complex lowland deciduous forest. Six sample plots in a pedunculate oak forest were measured in detail using conventional methods. Close-range remote sensing datasets used in this study represent solutions from low-cost sensors used for hand-held personal laser scanning (PLS_hh), unmanned–borne laser scanning (ULS) and unmanned aerial vehicle photogrammetry (UAV_image). Each tree in the sample plots was interactively measured directly from the point cloud, and correspondence of the field- and remote sensing measured trees was verified using tree positions collected during fieldwork. Cross-comparisons of different datasets were performed to evaluate the performances of different data sources in the tree height estimation with respect to crown class, tree height and species. All remote sensing data sources correlated well, e.g. biases between remote sensing sources were around ± 1%. The field-measured tree height in general correlated well with remote sensing data sources. The uncertainties and bias of the field measurements were dependent on the tree height and crown class. Field measurements tended to underestimate codominant and intermediate trees at the approximately 1 m magnitude, whilst remote sensing data sources were robust to crown classes. Low-cost ULS used in this study, and very likely in general, may not have enough penetration capability when measuring low and mostly occluded trees, causing missed treetops. PLS_hh gave tree height estimates closer to the real tree height than those derived from conventional field measurements for trees above 21 m height.

树高是森林清单中最重要的树属性之一。然而，使用常规的现场方法来测量树的高度是费力且费时的过程。尽管过去对促进树高测量非常感兴趣，但尚未对新的，即将到来的解决方案进行彻底研究。在这项研究中，我们调查了不同的近距离遥感方案对于复杂低地落叶林树高测量的适用性。使用常规方法详细测量了带花梗的橡树林中的六个样地。本研究中使用的近距离遥感数据集代表了用于手持式个人激光扫描（PLS _hh），无人驾驶激光扫描（ULS）和无人机航空摄影测量（UAV）的低成本传感器的解决方案_图片）。直接从点云中交互式地测量了样地中的每棵树，并使用在野外工作期间收集的树位置验证了实地和遥感测量的树的对应性。进行了不同数据集的交叉比较，以评估树冠高度，树高和树种方面的不同数据源在树高估计中的性能。所有遥感数据源之间的相关性都很好，例如，遥感源之间的偏差约为±1％。实地测得的树高通常与遥感数据源相关性很好。野外测量的不确定性和偏差取决于树木的高度和树冠等级。野外测量往往会低估大约1 m量级的优势树和中间树，遥感数据源对王冠等级具有鲁棒性。在这项研究中使用的低成本ULS（通常很可能）在测量低矮且大部分被遮挡的树木时可能没有足够的穿透能力，从而导致树梢遗漏。PLS_hh给出的树高估计值比从常规现场测量得出的高21 m以上的树时更接近真实树高。