2017年用于植物幼苗分类算法基准的公共图像数据库

植物幼苗数据集共包含407张 png 格式和不同大小的红绿蓝三色图像。图像涉及12种作物和杂草物种的植物幼苗中获取，在20天的生长期内多次拍摄。数据集创建者建造了一个便携式封闭框架，将相机固定在距土壤固定距离的位置表面并确保图像采集的光照条件均匀且可比较。每个获取的图像对应一个单一的植物物种，该数据集可以用于作物和杂草分类任务的基准。如果没有详细的杂草注释，该数据集不适合杂草分割或检测任务。https://vision.eng.au.dk/plant-seedlings-dataset/

The plant seedling dataset contains a total of 407 red-green-blue (RGB) color images in PNG format with varying sizes. These images were collected from plant seedlings of 12 crop and weed species, and were captured multiple times during a 20-day growth period. The dataset creators built a portable enclosed framework to fix the camera at a fixed distance above the soil surface, ensuring uniform and comparable lighting conditions for image acquisition. Each acquired image corresponds to a single plant species, and this dataset can serve as a benchmark for crop and weed classification tasks. This dataset is not suitable for weed segmentation or detection tasks without detailed weed annotations. https://vision.eng.au.dk/plant-seedlings-dataset/