This dataset contains 473 labeled guava fruit images that have been preprocessed with unsharp masking and CLAHE (contrast limited adaptive histogram equalization). The number of images after preprocessing increased to 3,784, and each image was preprocessed into a consistent RGB format of 512 x 512 pixels. These images are divided into 3 categories:

• Anthracnose
• Fruit Flies
• Healthy fruits

The images were collected in guava orchards in Rajshahi and Pabna, Bangladesh, during the fruit-ripening season in July, when the fruit is most vulnerable to disease.

Image details:

• Dimensions: 512 x 512 pixels
• Color Mode: RGB
• Format: PNG
• File size: 300–500 KB

This dataset is ideal for training and testing machine learning models designed to automatically detect common guava diseases, as well as for research on image classification and plant pathology image processing for agricultural applications.