MinneApple

MinneApple数据集由明尼苏达大学计算机科学与工程系创建，旨在推进果园环境中水果检测、分割和计数的技术。该数据集包含1000张高分辨率图像，涵盖多种苹果品种和不同生长阶段，共有超过41,000个标注对象实例。数据集通过多边形掩码为每个对象实例进行标注，以支持精确的对象检测、定位和分割。此外，还提供了基于补丁的密集水果计数数据。MinneApple数据集的应用领域主要集中在农业自动化，特别是苹果园的精准农业管理，如资源优化和收获决策支持。

The MinneApple dataset, developed by the Department of Computer Science and Engineering at the University of Minnesota, is designed to advance technologies for fruit detection, segmentation and counting in orchard environments. It contains 1,000 high-resolution images spanning multiple apple cultivars and diverse growth stages, with a total of over 41,000 annotated object instances. Each individual object instance is annotated using polygonal masks to support precise object detection, localization and segmentation. Additionally, patch-based dense fruit counting data is provided. The primary application domains of the MinneApple dataset focus on agricultural automation, particularly precision agricultural management in apple orchards, such as resource optimization and harvest decision support.