CWD30

CWD30是一个大规模、多样、全面的层次化数据集，专为精准农业中的作物杂草识别任务设计。它包含超过219,770张高分辨率图像，涵盖20种杂草和10种作物，涉及不同的生长阶段、多个视角和环境条件。这些图像从不同地理位置和季节的多样化农田中收集，确保了数据集的代表性。数据集的层次分类法支持细粒度分类，有助于开发更准确、稳健和可泛化的深度学习模型。

CWD30 is a large-scale, diverse and comprehensive hierarchical dataset tailored for crop and weed recognition tasks in precision agriculture. It comprises over 219,770 high-resolution images covering 20 weed species and 10 crop species, with a wide spectrum of growth stages, multiple viewing angles and varied environmental conditions. These images are acquired from diverse farmlands across different geographic locations and seasons, thus ensuring the representativeness of the dataset. The hierarchical taxonomy of the dataset enables fine-grained classification, facilitating the development of more accurate, robust and generalizable deep learning models.