HERMOS: An Annotated Image Dataset for Visual Detection of Grape Leaf Diseases

Vineyard powdery mildew (Uncinula necator), dead arm (Phomopsis viticola) and vineyard downy mildew (Plasmopara viticola) diseases are frequently seen in the vineyards in the Gediz River Basin, West Anatolia of Turkey and cause significant damage to the crop. These diseases can be detected early using artificial intelligence-based systems that can contribute to crop yields and also reduce the labor of the farmer and the amount of pesticides used. This article presents a dataset, for use in such AI-based systems. The dataset, namely Hermos, contains four classes of grapevine images; leaves with dead root, leaves with powdery mildew, leaves with downy mildew and healthy leaves. Diseased areas on the leaf pictures were labeled with the "bounding-box" method. The dataset contains a total of 914 images and 13,904 labels. Labels on the picture are stored in Pascal VOC format in an xml document with the same file name as the picture.土耳其安纳托利亚西部杰济德河（Gediz River）流域的葡萄园中，葡萄白粉病（Uncinula necator）、葡萄死臂病（Phomopsis viticola）与葡萄霜霉病（Plasmopara viticola）频发，此类病害会对葡萄作物造成严重损失。借助基于人工智能的系统可对上述病害实现早期检测，这不仅有助于提升作物产量，还能减轻农户的劳作负担并降低农药使用量。本文构建了一款可用于此类人工智能系统的数据集，该数据集命名为Hermos，包含四类葡萄叶片图像：表现出根腐症状的叶片、感染白粉病的叶片、感染霜霉病的叶片以及健康叶片。 叶片图像中的病斑区域采用边界框（bounding-box）标注法进行标注。本数据集总计包含914张图像与13904个标注框，图像的标注信息以帕斯卡VOC（Pascal VOC）格式存储于与图像同名的可扩展标记语言（XML）文档中。