Vegetables Image Dataset for Machine Applications

We developed the Vegetable Image Dataset to offer a collection of high-quality images featuring some of the most widely consumed and traded vegetables around the Pune. The dataset includes six types of vegetables: potato, chili, tomato, cucumber, beans, and okra. Each vegetable is further categorized into subclasses — potatoes are divided into three size-based classes (large, medium, and small) , while the other vegetables have two distinct varieties each (e.g., Chilies: Sitara, Jipoor, and Jwala; Tomatoes: Regular and Gaavran; Cucumbers: Regular and Gaavran; Beans: Long and Short; Okra: Long and Short) . This results in a total of 13 unique classes within the dataset . The images were taken under various lighting conditions — both natural and artificial — and against White backgrounds, including white, to ensure diversity and realism in the visual context. Given that the visual appearance of vegetables plays a significant role in their market value, this dataset supports research that evaluates vegetable quality through visual inspection . Although there are numerous datasets available for fruits and vegetables, many machine learning projects and applications still require a vegetable-specific dataset due to the unique nutritional importance and visual characteristics of vegetables . A robust dataset like this one enables machine learning models to achieve high accuracy in tasks such as classification and recognition . It's particularly useful for applications in research, education, and agriculture, including areas like detecting pest damage or monitoring quality degradation .