Fig 1.
The pipeline of the plant phenotyping system.
Fig 2.
A: System layout. B: The plant-phenotyping room. C,D: System configuration; the robotic arm moves the module from tray to tray and acquires top-view images.
Fig 3.
A: Top-view original tray images using color marker detection. B: The warped image based on the color markers.
Fig 4.
A: Estimation of tray edge coordinates. B: A cropped single-pot image.
Fig 5.
Top view of a tray showing pot labeling.
Fig 6.
Superpixel images.
Fig 7.
Plant-growth visualization; tracking plant area over time.
Fig 8.
The ground-truth image-creation user interface.
Fig 9.
A validation dataset (left) and a test dataset (right).
Fig 10.
The F1 scores of the three classifiers.
A: F1 scores of validation data sets. B: F1 scores of test data sets.
Fig 11.
Segmentation results after post-processing using three trained classifiers (Original, SVM, MLP, and RF).
Fig 12.
Precision-recall curves of the three classifiers.
A: Precision-recall curves of validation data sets. B: Precision-recall curves of test data sets.
Table 1.
Comparison of mean Average Precision.
Table 2.
Comparison of learning computation time.
Fig 13.
Plant-growth analysis; tracking plant area over time.