3D photogrammetric reconstructions of individual chickpea plants

3D point clouds and meshed models of individual chickpea plants were reconstructed using photogrammetry. This dataset serves as a companion to the article “Open source 3D phenotyping of chickpea plant architecture across plant development” uploaded as a preprint to BioRxiv (https://doi.org/10.1101/2020.09.08.288530).

Description of the dataset

Individual chickpea plants were imaged using a low-cost, automated turntable and camera set up. Dense point clouds were generated in VisualSFM, these point clouds were cleaned (denoised and non-plant points removed), scaled and then meshed using a ball-pivoting algorithm. More information on the imaging and reconstruction workflow can be found in our BioRxiv preprint (https://doi.org/10.1101/2020.09.08.288530). Scaled, clean point clouds and meshed models are provided in this dataset. All files are in the .PLY format and can be viewed using most 3D imaging software packages, including Meshlab.

Files are named using the following convention:
"GENOTYPE_PLANTIDNUMBER_GROUNDTRUTHINGORWEEKS_POINTCLOUDORMESHED"
Where:
GENOTYPE is the chickpea genotype.
PLANTIDNUMBER is the unique identifier for each plant.
GROUNDTRUTHINGORWEEKS refers to whether the plant was used for validation against ground truthing measurements or if growth was tracked after germination.
POINTCLOUDORMESHED refers to if the .ply file is a dense point cloud or a meshed 3D model.

Plant material

Three commercial Australian chickpea cultivars (Genesis Kalkee, PBA Hattrick and PBA Slasher) were grown in pots in a glasshouse and were imaged once per week for five weeks after germination. Three genotypes selected from local and international sources (ICC5878, SonSla and PUSA76) were grown in pots outdoors and were imaged once at five weeks post germination.