Assessing forest canopy dynamics is crucial for understanding the response of vegetation to environmental variability and change. While digital repeat photography is gaining increased attention for obtaining field phenology observations, colour indices derived from this method are often affected by leaf colour and actual canopy structure, complicating the physical interpretation of results. In addition, repeated photography requires power, storage capacity and remote data transfer, which are often limited in forest conditions. As an alternative, we tested a simple, cheap and fast solution to derive daily canopy structure observation from digital camera traps (CTs). Formerly deployed for wildlife monitoring, CTs are low-cost digital cameras designed for outdoor conditions and have low battery consumption, enable repeat acquisition, and often feature remote data transfer protocols.

The trial was performed in a deciduous oak stand, where continuous images were acquired over a 1-year period using the time-lapse feature of the CT. Daily time series of canopy structure attributes were derived from the collected images using simple and automated procedures. Results were validated against reference manual cover photography measurements. The daily time series of foliage cover and leaf area index were then used to derive phenological transition dates, which were compared against phenological observations obtained from satellite Sentinel-2 data. Results indicated that field and satellite data provided comparable accuracy in determining the start of season (SOS). Larger discrepancies were found in determining the end of season (EOS), which can be attributed to the low number of good quality autumn images available from the satellite data. We concluded that CT is a robust method, which is ideally suited for routine, continuous field monitoring of canopy attributes and phenology. While this method can be used for evaluating remote sensing observations, the combination of CTs with satellite data holds great potential for greater spatiotemporal coverage, from field to landscape scales.

评估森林冠层动态对于了解植被对环境变异和变化的响应至关重要。虽然数字重复摄影在获得野外物候观测方面越来越受到关注，但从这种方法得出的颜色指数通常会受到叶子颜色和实际冠层结构的影响，从而使结果的物理解释变得复杂。此外，重复拍摄需要电力、存储容量和远程数据传输，而这在森林条件下通常是有限的。作为替代方案，我们测试了一种简单、廉价且快速的解决方案，以从数码相机陷阱 (CT) 中获得日常冠层结构观察。CT 以前用于野生动物监测，是专为户外条件设计的低成本数码相机，电池消耗低，可重复采集，

该试验是在落叶橡木林中进行的，在那里使用 CT 的延时功能在 1 年的时间内获得了连续图像。使用简单和自动化的程序从收集的图像中导出冠层结构属性的每日时间序列。结果根据参考手动封面摄影测量进行了验证。然后使用树叶覆盖率和叶面积指数的每日时间序列来推导出物候过渡日期，并将其与从卫星 Sentinel-2 数据获得的物候观测结果进行比较。结果表明，现场和卫星数据在确定季节开始 (SOS) 方面提供了相当的准确性。在确定季末 (EOS) 时发现了更大的差异，这可归因于卫星数据提供的高质量秋季图像数量较少。我们得出结论，CT 是一种稳健的方法，非常适合对冠层属性和物候进行常规、连续的现场监测。虽然这种方法可用于评估遥感观测，但 CT 与卫星数据的结合具有更大的时空覆盖潜力，从田野到景观尺度。