Data for: Accurate Machine Learning-Based Germination Detection, Prediction and Quality Assessment of Various Seed Cultivars
Assessment of seed germination is an essential task for seed researchers to measure the quality and performance of seeds. Usually, seed assessments are done mainly manual, which is a cumbersome, time consuming and error-prone process. Classical image analyses methods are not well suited for large-scale germination experiments, because they often rely on manual adjustments of color-based thresholds. We here propose a machine learning approach using modern artificial neural networks for accurate seed germination detection for high-throughput seed germination experiments. This dataset contains captures of Germination Experiments of 824 Zea mays seeds, 811 Secale cereale seeds and 814 Pennisetum glaucum seeds. Approximately 10 seeds were placed in one petri dish and captured by a low-cost Raspberry Pi Camera Module (v2.1) in intervals of 30 minutes for ~ 2 days. All Images were annotated by Bounding Boxes containing their germination state (germinated/non-germinated). The code for running the models that are built with this data can be found on GitHub (https://github.com/grimmlab/GerminationPrediction).
Steps to reproduce
Use 85mm petri dishes with a black cloth piece for contrast. The petri dishes should be placed in a 4x3 grid. Place a low cost Raspberry Pi Camera Module (v2.1) above the petri dishes facing down to capture all 12 dishes as close as possible (~40 cm above the petri dish). Place ~10 seeds in each petri dish and start the germination experiment by spraying water on the seeds. Close the petri dishes with their lid and take a photograph every 30 minutes using the Raspberry Pi. The code for running the models that are built with this data can be found on GitHub (https://github.com/grimmlab/GerminationPrediction).