Guava Fruit Disease Dataset

Guava (Psidium guajava) is a vital crop in Bangladesh and South Asia. It is rich in vitamin C, fiber, and other nutrients. However, guava production is declining due to disease. Early detection of infections is essential to protect the harvest. With the rise of expert systems, automatic disease detection can mitigate these losses. This dataset focuses on guava fruit, providing images categorized into three classes: Anthracnose, Fruit Flies, and Healthy fruits. The images, collected from orchards in Rajshahi and Pabna, Bangladesh, were captured in July when fruits were ripening and were most prone to disease. The images were then verified by a plant pathologist. The dataset is intended to support research in image processing and machine learning, aiming to develop systems for early disease detection. Such systems can significantly reduce economic losses and improve guava yields.

The dataset includes 473 images of guava fruits. The original photos vary in size and format. Each was preprocessed to 512 x 512 pixels for consistency.

Image Information:
- Dimensions: 512 x 512 pixels
- Color Mode: RGB
- Format: PNG
- File Size: 300 - 500 KB

The dataset covers three main classes: Anthracnose, Fruit Flies, and Healthy fruits. These are common conditions in guava farming. Images underwent preprocessing steps such as unsharp masking and CLAHE. The preprocessed images are augmented to increase in number to 3,784 image data. The following tree shows the distribution of images across three subsets:

Dataset
│
├── train (2,647 images)
│ ├── Anthracnose
│ ├── Fruit Flies
│ └── Healthy fruits
│
├── test (382 images)
│ ├── Anthracnose
│ ├── Fruit Flies
│ └── Healthy fruits
│
└── val (775 images)
├── Anthracnose
├── Fruit Flies
└── Healthy fruits

石榴（Psidium guajava）在孟加拉国及南亚地区被视为一项至关重要的作物。其富含维生素C、纤维素及其他营养成分。然而，由于病害的影响，石榴的产量正逐渐下降。及早发现感染对于保护收成至关重要。随着专家系统的兴起，自动病害检测技术的应用有望减轻这些损失。本数据集聚焦于石榴果实，提供被分为三大类别的图像：炭疽病、果实蝇和健康果实。这些图像源自孟加拉国拉杰沙希和帕布纳的果园，拍摄于七月份果实成熟期，此时果实最易受到病害侵袭。图像随后经过植物病理学家的验证。本数据集旨在支持图像处理和机器学习领域的研究，旨在开发早期病害检测系统。此类系统可显著降低经济损失并提高石榴的产量。

数据集包含473张石榴果实的图像。原始照片尺寸和格式各异。每张照片均经过预处理，调整为512 x 512像素以保证一致性。

图像信息：
- 尺寸：512 x 512像素
- 颜色模式：RGB
- 格式：PNG
- 文件大小：300 - 500 KB

数据集涵盖了三大主要类别：炭疽病、果实蝇和健康果实。这些是石榴种植中的常见状况。图像经过预处理步骤，如非锐化掩蔽和对比度受限自适应直方图均衡化（CLAHE）。预处理后的图像经过增强，数量增至3,784个图像数据。以下树状图展示了图像在三个子集间的分布情况：

数据集
│
├── 训练集（2,647张图像）
│ ├── 炭疽病
│ ├── 果实蝇
│ └── 健康果实
│
├── 测试集（382张图像）
│ ├── 炭疽病
│ ├── 果实蝇
│ └── 健康果实
│
└── 验证集（775张图像）
├── 炭疽病
├── 果实蝇
└── 健康果实
}