ABSTRACT Phenotyping of genetic resources that are used for development of new improved and adapted forest reproductive materials is costly. Therefore, high-throughput-phenotyping from a standard equipped unmanned aerial vehicle (UAV) with a RGB (red, green, blue) camera was used to develop a canopy height model (CHM) of a dense field trial of 15-year-old trees of Norway spruce. The site represents a typical situation of genetic testing on abandoned farm field land, where spacing between trees often are around 1-1.5 m. The CHM was used to estimate total height, intra-annual height growth and phenology of individual trees. The estimates were based on individual tree grid cells assigned according to the original planting grid. Intra-annual growth and phenology could not be estimated accurately, whereas total tree height to the uppermost whorl was accurately estimated provided that the trees were correctly identified. However, the apices of suppressed trees were often misidentified due to overhanging branches from more dominating neighboring trees. The method represents a cost and time efficient alternative to manual height measurements of the intermediate and dominating trees in closed stands. The accuracy is sufficiently high for most breeding purposes when total tree height is the target trait.

中文翻译：

---

在森林遗传试验中使用无人机摄影测量数据：测量挪威云杉的树高、生长和物候（Picea abies L. Karst.）

摘要 用于开发新的改良和适应森林繁殖材料的遗传资源的表型分析是昂贵的。因此，使用配备 RGB（红、绿、蓝）相机的标准装备无人驾驶飞行器 (UAV) 的高通量表型来开发 15 岁密集田间试验的冠层高度模型 (CHM)。挪威云杉的树木。该站点代表了在废弃农田土地上进行基因检测的典型情况，其中树木之间的间距通常在 1-1.5 m 左右。CHM 用于估计个体树木的总高度、年内高度增长和物候。估计值基于根据原始种植网格分配的单个树网格单元。无法准确估计年内生长和物候，而如果树木被正确识别，则可以准确估计到最高轮的总树高。然而，由于更多支配性的相邻树木的悬垂树枝，被抑制的树木的顶端经常被错误识别。该方法代表了一种成本和时间效率高的替代方法，可以替代手动测量封闭林分中的中间树和主要树的高度。当总树高是目标性状时，对于大多数育种目的来说，准确度足够高。该方法代表了一种成本和时间效率高的替代方法，可以替代手动测量封闭林分中的中间树和主要树的高度。当总树高是目标性状时，对于大多数育种目的来说，准确度足够高。该方法代表了一种成本和时间效率高的替代方法，可以替代手动测量封闭林分中的中间树和主要树的高度。当总树高是目标性状时，对于大多数育种目的来说，准确度足够高。