A comprehensive image dataset for carambola leaf and fruit disease classification and quality assessment

The Carambola (Averrhoa carambola), also known as starfruit, is a tropical fruit with significant economic and nutritional value. The creation of a comprehensive Carambola Leaf and Fruit Dataset is highly essential for the advancement of automated disease classification and quality assessment using machine learning algorithms. The dataset is developed to support machine learning application and bridging the gap between computer vision and agricultural research to help farmers in minimizing financial losses and promote sustainable agricultural practices. The images are collected through direct field survey under diverse environmental conditions from various locations in Bangladesh between October 2024 and January 2025. The dataset comprises 2,618 original images, an equal number of processed images, and 15,000 augmented images generated from the original images. It is categorized into five distinct classes representing unique health conditions of carambola leaf and fruits including Healthy Leaves, Yellow Leaves, Insect Hole Leaves, Healthy Fruits, and Unhealthy Fruits. This dataset supports sustainable agriculture by enabling early disease identification, reduced chemical usage, and improved crop management while minimizing economic losses for farmers. It serves as a valuable resource for the future research in machine learning based plant health monitoring and quality assessment.