skip to main content

Title: Semantic Segmentation of Sorghum Using Hyperspectral Data Identifies Genetic Associations
This study describes the evaluation of a range of approaches to semantic segmentation of hyperspectral images of sorghum plants, classifying each pixel as either nonplant or belonging to one of the three organ types (leaf, stalk, panicle). While many current methods for segmentation focus on separating plant pixels from background, organ-specific segmentation makes it feasible to measure a wider range of plant properties. Manually scored training data for a set of hyperspectral images collected from a sorghum association population was used to train and evaluate a set of supervised classification models. Many algorithms show acceptable accuracy for this classification task. Algorithms trained on sorghum data are able to accurately classify maize leaves and stalks, but fail to accurately classify maize reproductive organs which are not directly equivalent to sorghum panicles. Trait measurements extracted from semantic segmentation of sorghum organs can be used to identify both genes known to be controlling variation in a previously measured phenotypes (e.g., panicle size and plant height) as well as identify signals for genes controlling traits not previously quantified in this population (e.g., stalk/leaf ratio). Organ level semantic segmentation provides opportunities to identify genes controlling variation in a wide range of morphological phenotypes in sorghum, maize, and other related grain crops.  more » « less
Award ID(s):
Author(s) / Creator(s):
; ; ; ; ;
Date Published:
Journal Name:
Plant Phenomics
Page Range / eLocation ID:
1 to 11
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract Most semantic segmentation approaches of big data hyperspectral images use and require preprocessing steps in the form of patching to accurately classify diversified land cover in remotely sensed images. These approaches use patching to incorporate the rich spatial neighborhood information in images and exploit the simplicity and segmentability of the most common datasets. In contrast, most landmasses in the world consist of overlapping and diffused classes, making neighborhood information weaker than what is seen in common datasets. To combat this common issue and generalize the segmentation models to more complex and diverse hyperspectral datasets, in this work, we propose a novel flagship model: Clustering Ensemble U-Net. Our model uses the ensemble method to combine spectral information extracted from convolutional neural network training on a cluster of landscape pixels. Our model outperforms existing state-of-the-art hyperspectral semantic segmentation methods and gets competitive performance with and without patching when compared to baseline models. We highlight our model’s high performance across six popular hyperspectral datasets including Kennedy Space Center, Houston, and Indian Pines, then compare them to current top-performing models. 
    more » « less
  2. Premise

    Maize yields have significantly increased over the past half‐century owing to advances in breeding and agronomic practices. Plants have been grown in increasingly higher densities due to changes in plant architecture resulting in plants with more upright leaves, which allows more efficient light interception for photosynthesis. Natural variation for leaf angle has been identified in maize and sorghum using multiple mapping populations. However, conventional phenotyping techniques for leaf angle are low throughput and labor intensive, and therefore hinder a mechanistic understanding of how the leaf angle of individual leaves changes over time in response to the environment.


    High‐throughput time series image data from water‐deprived maize (Zea mayssubsp.mays) and sorghum (Sorghum bicolor) were obtained using battery‐powered time‐lapse cameras. A MATLAB‐based image processing framework, Leaf Angle eXtractor (LAX), was developed to extract and quantify leaf angles from images of maize and sorghum plants under drought conditions.


    Leaf angle measurements showed differences in leaf responses to drought in maize and sorghum. Tracking leaf angle changes at intervals as short as one minute enabled distinguishing leaves that showed signs of wilting under water deprivation from other leaves on the same plant that did not show wilting during the same time period.


    Automating leaf angle measurements using LAX makes it feasible to perform large‐scale experiments to evaluate, understand, and exploit the spatial and temporal variations in plant response to water limitations.

    more » « less
  3. Abstract

    Grass leaves develop from a ring of primordial initial cells within the periphery of the shoot apical meristem, a pool of organogenic stem cells that generates all of the organs of the plant shoot. At maturity, the grass leaf is a flattened, strap-like organ comprising a proximal supportive sheath surrounding the stem and a distal photosynthetic blade. The sheath and blade are partitioned by a hinge-like auricle and the ligule, a fringe of epidermally derived tissue that grows from the adaxial (top) leaf surface. Together, the ligule and auricle comprise morphological novelties that are specific to grass leaves. Understanding how the planar outgrowth of grass leaves and their adjoining ligules is genetically controlled can yield insight into their evolutionary origins. Here we use single-cell RNA-sequencing analyses to identify a ‘rim’ cell type present at the margins of maize leaf primordia. Cells in the leaf rim have a distinctive identity and share transcriptional signatures with proliferating ligule cells, suggesting that a shared developmental genetic programme patterns both leaves and ligules. Moreover, we show that rim function is regulated by genetically redundant Wuschel-like homeobox3 (WOX3) transcription factors. Higher-order mutations in maizeWox3genes greatly reduce leaf width and disrupt ligule outgrowth and patterning. Together, these findings illustrate the generalizable use of a rim domain during planar growth of maize leaves and ligules, and suggest a parsimonious model for the homology of the grass ligule as a distal extension of the leaf sheath margin.

    more » « less
  4. Plant communities are composed of complex phenotypes that not only differ among taxonomic groups and habitats but also change over time within a species. Restoration projects (e.g. translocations and reseeding) can introduce new functional variation in plants, which further diversifies phenotypes and complicates our ability to identify locally adaptive phenotypes for future restoration. Near‐infrared spectroscopy (NIRS) offers one approach to detect the chemical phenotypes that differentiate plant species, populations, and phenological states of individual plants over time. We use sagebrush (Artemisiaspp.) as a case study to test the accuracy by which NIRS can classify variation within taxonomy and phenology of a plant that is extensively managed and restored. Our results demonstrated that NIRS can accurately classify species of sagebrush within a study site (75–96%), populations of sagebrush within a subspecies (99%), annual phenology within a population (>99%), and seasonal phenology within individual plants (>97%). Low classification accuracy by NIRS in some sites may reflect heterogeneity associated with natural hybridization, translocation of nonlocal seed sources from past restoration, or complex gene‐by‐environment interactions. Advances in our ability to detect and interpret spectral signals from plants may improve both the selection of seed sources for targeted conservation and the capacity to monitor long‐term changes in vegetation.

    more » « less
  5. Virtually all land plants are coated in a cuticle, a waxy polyester that prevents nonstomatal water loss and is important for heat and drought tolerance. Here, we describe a likely genetic basis for a divergence in cuticular wax chemistry between Sorghum bicolor , a drought tolerant crop widely cultivated in hot climates, and its close relative Zea mays (maize). Combining chemical analyses, heterologous expression, and comparative genomics, we reveal that: 1) sorghum and maize leaf waxes are similar at the juvenile stage but, after the juvenile-to-adult transition, sorghum leaf waxes are rich in triterpenoids that are absent from maize; 2) biosynthesis of the majority of sorghum leaf triterpenoids is mediated by a gene that maize and sorghum both inherited from a common ancestor but that is only functionally maintained in sorghum; and 3) sorghum leaf triterpenoids accumulate in a spatial pattern that was previously shown to strengthen the cuticle and decrease water loss at high temperatures. These findings uncover the possibility for resurrection of a cuticular triterpenoid-synthesizing gene in maize that could create a more heat-tolerant water barrier on the plant’s leaf surfaces. They also provide a fundamental understanding of sorghum leaf waxes that will inform efforts to divert surface carbon to intracellular storage for bioenergy and bioproduct innovations. 
    more » « less