Comprehensive dataset on ripening stages of strawberries and avocados: From unripe to rotten.

Abstract

This paper presents a novel and innovative 14,630 fruit images dataset, consisting of 1333 original images and the remaining augmented images for strawberry and avocado fruits. The dataset records the growth of strawberries and avocados in four different stages: unripe, partially ripe, ripe, and rotten. Though the fruit ripening process is commonly known, a lack of systematic datasets to show the fruit changing from an unripe state to a rotting state was prevalent for the two fruits in question. Over two months, the dataset was collected through rigorous tracking to effectively provide a measure of each of the fruits' conditions. The fruits were obtained from Mahabaleshwar farms in Maharashtra, India, as well as from local markets in Maharashtra and Pune. The fruits were monitored continuously from the time of harvesting, and all observed changes were carefully recorded. The uniqueness of this dataset is that it covers both strawberries and avocados, which have different patterns of ripening and are highly commercially valuable. The images were annotated using the online annotation tool - makesense.ai, with a total of 1499 bounding boxes for each fruit. By encompassing these two diverse fruit types, the dataset provides a valuable resource for researchers, agriculturalists, and food scientists to investigate and compare the ripening behaviours of different fruit species.

Plant phenotyping relevance

イチゴとアボカドの果実画像を用いて、未熟から腐敗までの可視的な成熟・状態を体系的に記録し、注釈付きデータセットとして提供しているため、植物器官の状態を対象とする画像ベースのフェノタイピングデータセットに該当する。

Code and data availability

The paper is a Data in Brief article describing a public Mendeley Data repository containing the authors' own fruit image dataset (14,630 strawberry/avocado images with YOLO bounding-box annotations across ripening stages), which directly constitutes the paper's phenotyping measurements. No analysis code or trained模型s是