Real-world Tomato Image Dataset for Deep Learning and Computer Vision Applications Involving Precision Agriculture

Published: 09-11-2020| Version 1 | DOI: 10.17632/9zyvdgp83m.1
RajinderKumar M Math,
Dr. Nagaraj V. Dharwadkar


The dataset consists of high resolution real images of tomato fruit (vegetable) which were taken at various stages of tomato growth starting from flowering all the way to harvesting stage over a period of 1 year. The dataset was created keeping in mind the real-time scenario that helps in obtaining good generalisation capability for the Deep Learning model or any other model. The dataset would help the researchers working towards agriculture domain utilising the capabilities of Deep Learning and Computer Vision. The images were collected at different times (morning, evening and night) with various backgrounds. The dataset consists of two folders of tomato images, the first folder consists of 1114 mixed images of tomato (growth stages) having a resolution of 1846x4000 while the second folder consists of 636 mixed tomato images having a resolution of 4000x1660. The images were captured by using Redmi K20 Pro phone that uses a 48-megapixel Sony IMX586 sensor with an f/1.75 aperture and 1.6 micron pixels.