Plant Phenomics / 2020 / Article / Fig 2

Research Article

A Deep Learning-Based Phenotypic Analysis of Rice Root Distribution from Field Images

Figure 2

Calculating root distribution parameters. (a) The trench profile image was normalized as 60 cm in depth by 60 cm in width, with 256 dots per 10 cm. The image was folded in half and skeletonized to calculate the root distribution parameters. (b) A schematic diagram of how to calculate the root distribution parameters. In the case of Depth50, the cumulative probability along the vertical axis was calculated; the depth, with its cumulative probability of 0.5, was defined as Depth50. In the case of Width50, the cumulative probability along the horizontal axis was calculated; the width, with its cumulative probability of 0.5, was defined as Width50.
(a)
(b)