Land fragmentation and small plots are the main features of the rural environment of Galicia (NW Spain). Smallholding limits land use management, representing a drawback in local forest planning. This study analyzes the potential use of multitemporal Sentinel-2 images to detect and control forest cuts in very small pine and eucalyptus plots located in southern Galicia. The proposed approach is based on the analysis of Sentinel-2 NDVI time series in 4231 plots smaller than 3 ha (average 0.46 ha). The methodology allowed us to detect cuts, allocate cut dates and quantify plot areas due to different cutting cycles in an uneven-aged stand. An accuracy of approximately 95% was achieved when the whole plot was cut, with an 81% accuracy for partial cuts. The main difficulty in detecting and dating cuts was related to cloud cover, which affected the multitemporal analysis. In conclusion, the proposed methodology provides an accurate estimation of cutting date and area, helping to improve the monitoring system in sustainable forest certifications to ensure compliance with forest management plans.

中文翻译：

---

小规模森林砍伐控制的多时态 Sentinel-2 数据分析

土地碎片化和小地块是加利西亚（西班牙西北部）农村环境的主要特征。小规模经营限制了土地使用管理，这是当地森林规划的一个缺陷。本研究分析了多时相 Sentinel-2 图像在位于加利西亚南部的非常小的松树和桉树地块中检测和控制森林砍伐的潜在用途。建议的方法基于对小于 3 公顷（平均 0.46 公顷）的 4231 个地块中的 Sentinel-2 NDVI 时间序列的分析。该方法使我们能够检测切割、分配切割日期并量化由于不均匀老化林分中的不同切割周期而导致的小区面积。切割整块地块的准确度约为 95%，部分切割的准确度为 81%。检测和测年的主要困难与云量有关，这影响了多时态分析。总之，所提出的方法提供了对砍伐日期和面积的准确估计，有助于改进可持续森林认证的监测系统，以确保符合森林管理计划。