2019年早期作物和杂草图像数据集

早期作物杂草数据集能够将预训练的深度学习模型应用于作物和杂草识别。该数据集针对处于早期生长阶段的两种杂草，有3-4片叶子的黑色茄属植物和天鹅绒，还包含两种作物，番茄和棉花。图像数据是使用 红绿蓝 相机在自然光条件下于不同场地位置拍摄。生成的黑色茄属植物、天鹅绒、番茄和棉花数据集分别包含123、130、54和201 张图像，均为 jpg 格式和4256×2832像素。采集到的图片被归入表示相应植物类别的文件夹中。由于每个图像包含单一植物物种，该数据集不适合作物杂草语义分割和定位任务。https://github.com/AUAgroup/early-crop-weed

The Early Crop-Weed Dataset facilitates the deployment of pre-trained deep learning models for crop and weed recognition. This dataset covers two weed species, black nightshade (with 3–4 true leaves) and velvetleaf, as well as two crop species, tomato and cotton, during their early growth stages. Image data was captured using RGB cameras under natural light conditions at multiple field sites. The datasets for black nightshade, velvetleaf, tomato, and cotton respectively contain 123, 130, 54, and 201 images, all in JPG format with a resolution of 4256×2832 pixels. Collected images are organized into folders corresponding to their respective plant categories. Since each image contains only a single plant species, this dataset is not applicable to crop-weed semantic segmentation and localization tasks. https://github.com/AUAgroup/early-crop-weed