Niedz Randall P, Bowman Kim D
Agricultural Research Service, U.S. Horticultural Research Laboratory, 2001 South Rock Road, Ft. Pierce, FL, USA.
Data Brief. 2024 Dec 5;58:111206. doi: 10.1016/j.dib.2024.111206. eCollection 2025 Feb.
The data are aerial images and ground tree measurement data of 3 citrus rootstock trials. Developing new citrus rootstock varieties requires field trials to test to identify selections with improved horticultural performance. A bud from a scion variety is grafted onto the rootstock and grown in a nursery until the grafted plant is ready to be planted in the field, which is in about one year. Trees in the field are assessed each year by measuring height, canopy diameter in 2 dimensions, overall health, and fruit number and quality factors when the trees begin to have a significant crop (∼3 years). Data collection of each tree is done manually. The image and ground data sets are of 3 rootstock trials that includes a 3-year-old Bingo mandarin hybrid trial of 206 trees, a 6-year-old Valencia orange trial of 643 trees, and a 7-year-old Valencia orange trials of 648 trees. Data for each trial includes aerial images and ground data of height, canopy diameters, and an overall health rating. The combination of ground validated measures and aerial images make this data set useful for building AI-based aerial image data collection applications. The data will be useful for 1) visualizing the effects of different rootstock selections and varieties on scion growth, effects that may not be fully captured with single measure metrics; and 2) development of image analysis applications and segmentation algorithms that can extract data from the images that are suitable for replacing some or all the ground measures.
这些数据是3个柑橘砧木试验的航空图像和地面树木测量数据。培育新的柑橘砧木品种需要进行田间试验，以测试并识别具有改良园艺性能的选择。将接穗品种的一个芽嫁接到砧木上，在苗圃中生长，直到嫁接植株准备好移栽到田间，这大约需要一年时间。当树木开始大量结果（约3年）时，每年通过测量高度、二维树冠直径、整体健康状况以及果实数量和品质因素来评估田间的树木。每棵树的数据收集都是手动完成的。图像和地面数据集来自3个砧木试验，包括一项有206棵树的3年生宾果杂交柑橘试验、一项有643棵树的6年生巴伦西亚橙试验以及一项有648棵树的7年生巴伦西亚橙试验。每个试验的数据包括航空图像以及高度、树冠直径和整体健康评级的地面数据。地面验证测量和航空图像的结合使得该数据集对于构建基于人工智能的航空图像数据收集应用很有用。这些数据将有助于：1）直观呈现不同砧木选择和品种对接穗生长的影响，这些影响可能无法通过单一测量指标完全体现；2）开发能够从图像中提取适合替代部分或全部地面测量数据的图像分析应用和分割算法。
