Sony/Hokkaido_Agriculture_Image_Dataset

北海道农业图像数据集是由北海道大学和索尼集团合作构建的，包含大约2000张来自北海道不同农业用地的图片，涵盖葡萄、小麦、洋葱、苹果和黑加仑五种作物。这些图片覆盖了不同的生长阶段、光照条件和视角，以支持在实际农业环境中的图像识别和机器学习研究。数据集包括针对收获等应用的实例分割数据，以及用于准确确定作物数量的对象检测数据。所有图像的标注格式均为COCO格式。

The Hokkaido Agriculture Image Dataset, constructed in collaboration between Hokkaido University and Sony Group Corporation, consists of approximately 2,000 images covering five crops: grapes, wheat, onions, apples, and haskap. These images are collected from various agricultural lands in Hokkaido and cover different growth stages, lighting conditions, and perspectives to support image recognition and machine learning research in actual agricultural environments. The dataset includes instance segmentation data for applications such as harvesting and object detection data for accurately determining crop quantities. All images are annotated in the COCO format.