花生根微根管图像数据集和柳枝稷根微根管图像数据集
收藏资源简介:
本研究构建了两个微根管根图像数据集:一个包含17,550张花生根RGB图像,另一个包含28张柳枝稷根RGB图像。两个数据集均配有人工标注的地面实况掩码,用于指示每张图像中根的位置。花生根数据集在2016年生长季节在佛罗里达州Citra的植物科学研究与教育单位(PSREU)的田间试验中收集。柳枝稷根数据集使用CI-602 insitu根成像仪在伊利诺伊州Batavia的费米实验室国家环境研究公园的一个两年生柳枝稷田中收集。这些数据集用于研究转移学习技术在有限数据集上的应用,特别是在植物根系分割任务中,旨在提高数据收集和后处理的效率,解决全球粮食、资源和气候问题。
This study constructed two minirhizotron root image datasets: one containing 17,550 RGB images of peanut roots, and the other containing 28 RGB images of switchgrass roots. Both datasets are paired with manually annotated ground truth masks that indicate the positions of roots in each image. The peanut root dataset was collected during the 2016 growing season from a field experiment conducted at the Plant Science Research and Education Unit (PSREU) in Citra, Florida, USA. The switchgrass root dataset was collected using a CI-602 in-situ root imager from a two-year-old switchgrass field at the Fermilab National Environmental Research Park in Batavia, Illinois, USA. These datasets are used to study the application of transfer learning techniques on limited datasets, particularly in plant root segmentation tasks, aiming to improve the efficiency of data collection and post-processing, and address global food, resource, and climate issues.




