{Reference Type}: Journal Article {Title}: An instance segmentation dataset of cabbages over the whole growing season for UAV imagery. {Author}: Yokoyama Y;Matsui T;Tanaka TST; {Journal}: Data Brief {Volume}: 55 {Issue}: 0 {Year}: 2024 Aug 暂无{DOI}: 10.1016/j.dib.2024.110699 {Abstract}: Crop growth monitoring is essential for both crop and supply chain management. Conventional manual sampling is not feasible for assessing the spatial variability of crop growth within an entire field or across all fields. Meanwhile, UAV-based remote sensing enables the efficient and nondestructive investigation of crop growth. A variety of crop-specific training image datasets are needed to detect crops from UAV imagery using a deep learning model. Specifically, the training dataset of cabbage is limited. This data article includes annotated cabbage images in the fields to recognize cabbages using machine learning models. This dataset contains 458 images with 17,621 annotated cabbages. Image sizes are approximately 500 to 1000 pixel squares. Since these cabbage images were collected from different cultivars during the whole growing season over the years, deep learning models trained with this dataset will be able to recognize a wide variety of cabbage shapes. In the future, this dataset can be used not only in UAVs but also in land-based robot applications for crop sensing or associated plant-specific management.