XingjianLi/tomatotest

该数据集包含21,384对2448x2048分辨率的合成图像，每对图像包括左右RGBD图像和左图像的全景分割标签。数据集支持图像分割和深度估计任务，主要用于农业研究机器人平台的密集植物病害检测。数据集的创建基于PlantVillage数据集，并进行了进一步处理以对齐健康叶片的颜色。

This dataset consists of 21,384 pairs of synthetic images, each pair including left and right RGBD images and panoptic segmentation labels for the left image. The dataset supports image segmentation and depth estimation tasks, with English as the language. Data instances include six types of images, with detailed descriptions of each images characteristics in the data fields. The data split is 80/20. The creation motivation of the dataset is to provide a dense plant disease detection dataset for an agricultural research robotics platform. The source data comes from the PlantVillage dataset and has been further processed. The annotation process is automatically generated through textures in the simulation.