Pomegranate Fruit Diseases Dataset for Deep Learning Models

Pomegranate Fruit Diseases datasets should be efficient and clean for creating accurate and consistent machine-learning models and minimizing misclassification in real-time scenarios. The publicly available, standardized datasets for the categorization of pomegranate fruit illnesses currently used in agriculture are insufficiently precise and comprehensive to effectively train the models. Our main objective for the current project is to produce a publicly accessible dataset of pomegranate fruits and an image dataset of pomegranate fruits with various diseases that is ready for use to address this issue. Five different pomegranate fruit diseases—from areas like Ballari, Bangalore, Bagalkote, etc.—have been compiled by us. The photos were shot in July and October of 2023. The collection includes 5099 labeled and categorized images of healthy and diseased pomegranate fruit, divided into five categories: Healthy, Bacterial blight, Anthracnose, Cercospora fruit spot, and Alternaria fruit spot. The dataset consists of five folders named by related medical conditions. This dataset may help identify pomegranate diseases in other countries and boosting pomegranate yield output.

石榴果实病害数据集需具备高效性与干净性，以构建精准且一致的机器学习模型，并在实时应用场景中降低分类误判率。当前农业领域用于石榴果实病害分类的公开标准化数据集，在精准度与全面性上均存在不足，无法有效支撑模型训练。本项目的核心目标正是针对这一痛点，构建可公开获取的石榴果实数据集，以及可供直接使用的多病害石榴果实图像数据集。我们从巴拉里（Ballari）、班加罗尔（Bangalore）、巴加果特（Bagalkote）等区域采集了5种不同的石榴果实病害样本，图像采集时间为2023年7月与10月。本数据集共包含5099张经过标注与分类的健康与病害石榴果实图像，划分为5个类别：健康（Healthy）、细菌性疫病（Bacterial blight）、炭疽病（Anthracnose）、尾孢果斑病（Cercospora fruit spot）以及链格孢果斑病（Alternaria fruit spot）。该数据集按照对应病害类别分为5个文件夹。本数据集可助力其他国家开展石榴病害识别工作，进而提升石榴总产量。