2018年基于可变光照条件下RGB图像的杂草检测数据集

萝卜-杂草数据集包含在胡萝卜田中收集的图像。图像在自然光条件下使用低成本手机摄像头拍摄。它也是一个小规模数据集，由39张红绿蓝三色图像组成，这些图像为jpg格式，分辨率为1296×966 像素，此外还有两组分别定义植被掩膜和作物、杂草和土壤背景的像素级注释的图像。该数据集已用于结合随机森林的纹理特征进行作物和杂草识别。https://github.com/lameski/rgbweeddetection

The Radish-Weed Dataset comprises images collected in carrot fields. All images were acquired under natural illumination using low-cost smartphone cameras. This is a small-scale dataset containing 39 RGB images in JPEG format with a resolution of 1296×966 pixels. Additionally, two sets of pixel-level annotated images are included, which respectively define vegetation masks and the backgrounds of crops, weeds and soil. This dataset has been utilized for crop and weed recognition by integrating texture features with random forests. https://github.com/lameski/rgbweeddetection