Voxel carving‐based 3D reconstruction of sorghum identifies genetic determinants of light interception efficiency

Abstract

Changes in canopy architecture traits have been shown to contribute to yield increases. Optimizing both light interception and light interception efficiency of agricultural crop canopies will be essential to meeting the growing food needs. Canopy architecture is inherently three‐dimensional (3D), but many approaches to measuring canopy architecture component traits treat the canopy as a two‐dimensional (2D) structure to make large scale measurement, selective breeding, and gene identification logistically feasible. We develop a high throughput voxel carving strategy to reconstruct 3D representations of sorghum from a small number of RGB photos. Our approach builds on the voxel carving algorithm to allow for fully automatic reconstruction of hundreds of plants. It was employed to generate 3D reconstructions of individual plants within a sorghum association population at the late vegetative stage of development. Light interception parameters estimated from these reconstructions enabled the identification of known and previously unreported loci controlling light interception efficiency in sorghum. The approach is generalizable and scalable, and it enables 3D reconstructions from existing plant high throughput phenotyping datasets. We also propose a set of best practices to increase 3D reconstructions’ accuracy.

High-Throughput Corn Image Segmentation and Trait Extraction Using Chlorophyll Fluorescence Images

“…Also, Choudhury et al and Khan et al used background removal, which depends on the light intensity and regular calibration of the system [ 42 , 47 ]. A convolutional neural network, a more advanced techniques, was leveraged for segmentation [ 48 ]; however, it was more practical for the AAPF to adopt the CFI segmentation. Regarding stem coordinate estimation, our approach was different and led to diverse results with a different interpretation compared to these researchers' work when using our data set since we estimated the stem path from surface to the stalk node with the youngest fully developed leaf (top node).…”

Automation of Leaf Counting in Maize and Sorghum Using Deep Learning

“…Leaf occlusion is therefore likely to introduce heritable genetic error (Liang et al, 2017) if counting-by-detection is employed to phenotype maize plants in quantitative genetic or plant breeding projects. Transitions to methods which incorporate 3D information (McCormick et al, 2016;Xiang et al, 2019;Gaillard et al, 2020) and/or the development of approaches to integrate information on leaf tips detected in multiple images of the same plant taken from different angles will likely be ultimately needed.…”

“…The target photoperiod was 14:10 with supplementary light provided by light-emitting diode (LED) growth lamps from 07:00 to 21:00 each day. The target temperature of the growth facility was between 24-26 • C. After growing in the greenhouse for 28 days in the first experiment and 38 days in the second experiment, all the plants were moved to a high throughput phenotyping facility equipped with a plant conveying system and different kinds of imaging chambers (Ge et al, 2016;Gaillard et al, 2020). The conveyor belt transferred each pot to a set of imaging chambers once every two or three days for imaging, and to a watering station each day.…”

3D reconstruction identifies loci linked to variation in angle of individual sorghum leaves

“…A set of 366 individual plants from 236 genotypes of the sorghum association panel [23] were grown and imaged using the University of Nebraska-Lincoln's automated phenotyping facility [32]. Growth conditions and details of the image acquisition were previously described in [41]. RGB images were taken on April 11th (47 days after planting), 13th (49 days after planting) and 16th (52 days after planting) of 2018, using a Basler pia2400-17gc camera with a c6z1218m3-5 Pentax TV zoom lens.…”

“…3D volumentric grid was reconstructed for each plant on each day using an extended voxel carving algorithm [41]. The algorithm employed segmented images from each view to carve a voxel grid with a resolution of 512 3 voxels.…”