Dataset of tomato plants growth observations obtained from multiple sources in a production-like setting

• Autor(a) principal: Oliveira, Monique
• Data de Publicação: 2023
• Outros Autores: Amaro, Rafaella, Pescarini, Henrique, Rodrigues, Luiz Henrique
• Tipo de documento: preprint
• Idioma: eng
• Título da fonte: SciELO Preprints
• Texto Completo: https://preprints.scielo.org/index.php/scielo/preprint/view/7667
• topic: computer vision; Protected growth; Image segmentation; Greenhouse; Digital agriculture

This dataset contains observations of tomato growth in a production-like setting, at research greenhouses. Two plants in each of three growth cycles were continuously monitored and pictures were taken every other day from above and from a side view while a weighting system was used to record changes in weight of plant and water in the substrate. Other plants in the environment were subjected to destructive analysis in general every two weeks to quantify aspects of growth that required destructive measurements, such as dry weight and plant leaf area, and these records are also included in the dataset, including the scans of digitized leaves. Plant samples destined to destructive measurements also had their pictures taken before removal. In total, 618 photos of monitored and removed plants were annotated, and masks of leaf, fruit and mature fruit are also provided. The dataset also includes measurements of photosynthetically active radiation and air temperature recorded inside the greenhouses by two sets of different sensors during the growth cycles. The dataset allows for applications regarding growth monitoring and computer vision tasks.