WheatSpikeNet: an improved wheat spike segmentation model for accurate estimation from field imaging.

Phenotyping is used in plant breeding to identify genotypes with desirable characteristics, such as drought tolerance, disease resistance, and high-yield potentials. It may also be used to evaluate the effect of environmental circumstances, such as drought, heat, and salt, on plant growth and development. Wheat spike density measure is one of the most important agronomic factors relating to wheat phenotyping. Nonetheless, due to the diversity of wheat field environments, fast and accurate identification for counting wheat spikes remains one of the challenges. This study proposes a meticulously curated and annotated dataset, named as SPIKE-segm, taken from the publicly accessible SPIKE dataset, and an optimal instance segmentation approach named as WheatSpikeNet for segmenting and counting wheat spikes from field imagery. The proposed method is based on the well-known Cascade Mask RCNN architecture with model enhancements and hyperparameter tuning to provide state-of-the-art detection and segmentation performance. A comprehensive ablation analysis incorporating many architectural components of the model was performed to determine the most efficient version. In addition, the model's hyperparameters were fine-tuned by conducting several empirical tests. ResNet50 with Deformable Convolution Network (DCN) as the backbone architecture for feature extraction, Generic RoI Extractor (GRoIE) for RoI pooling, and Side Aware Boundary Localization (SABL) for wheat spike localization comprises the final instance segmentation model. With bbox and mask mean average precision (mAP) scores of 0.9303 and 0.9416, respectively, on the test set, the proposed model achieved superior performance on the challenging SPIKE datasets. Furthermore, in comparison with other existing state-of-the-art methods, the proposed model achieved up to a 0.41% improvement of mAP in spike detection and a significant improvement of 3.46% of mAP in the segmentation tasks that will lead us to an appropriate yield estimation from wheat plants.

Modelling of PFAS-surface interactions: Effect of surface charge and solution ions.

PER-: and poly-fluoroalkyl substances (PFAS) are a class of substances of increasing concern as environmental contaminants. The interactions between PFAS and surfaces play an important role in PFAS transport and remediation. Previous studies have found PFAS adsorption to be dependent upon properties including pH, organic matter and particle size, along with PFAS functional group and carbon chain length. It is hypothesised that a theoretical examination of PFAS-surface interactions, via Monte Carlo molecular simulation, would show differences resulting from changes in surface charge, H+, OH-, Ca2+ concentrations and PFAS carbon chain length. Monte Carlo molecular simulations of perfluorooctane and perfluorobutane sulfonic acids interacting with a graphite surface in an aqueous medium were performed. Variations in surface charge, H+, OH- and Ca2+ concentrations were made. The distance-dependent density of molecules from the surface was analysed as a proxy for PFAS adsorption to the surface. Simulation results showed differences in surface behaviour that depended on surface charge, H+, OH- and Ca2+ concentrations, along with carbon chain length, with surface charge playing the most prominent role in controlling PFAS adsorption. For negatively charged surfaces, adsorption due to divalent cation bridging was observed in Ca2+ solutions. Modelling, such as in this study, of the thermodynamic equilibrium behaviour of low concentrations of molecules, in scenarios where both adsorption and mobility of PFAS occur, can aid in the design and testing of sorptive surfaces for amendment-based PFAS remediation.

A Whole Leaf Comparative Study of Stomatal Conductance Models.

We employed a detailed whole leaf hydraulic model to study the local operation of three stomatal conductance models distributed on the scale of a whole leaf. We quantified the behavior of these models by examining the leaf-area distributions of photosynthesis, transpiration, stomatal conductance, and guard cell turgor pressure. We gauged the models' local responses to changes in environmental conditions of carbon dioxide concentration, relative humidity, and light irradiance. We found that a stomatal conductance model that includes mechanical processes dependent on local variables predicts a spatial variation of physiological activity across the leaf: the leaf functions of photosynthesis and transpiration are not uniformly operative even when external conditions are uniform. The gradient pattern of hydraulic pressure which is needed to produce transpiration from the whole leaf is derived from the gradient patterns of turgor pressures of guard cells and epidermal cells and consequently leads to nonuniform spatial distribution patterns of transpiration and photosynthesis via the mechanical stomatal model. Our simulation experiments, comparing the predictions of two versions of a mechanical stomatal conductance model, suggest that leaves exhibit a more complex spatial distribution pattern of both photosynthesis and transpiration rate and more complex dependencies on environmental conditions when a non-linear relationship between the stomatal aperture and guard cell and epidermal cell turgor pressures is implemented. Our model studies offer a deeper understanding of the mechanism of stomatal conductance and point to possible future experimental measurements seeking to quantify the spatial distributions of several physiological activities taking place over a whole leaf.

Interrogating the relationship between the microstructure of amphiphilic poly(ethylene glycol-b-caprolactone) copolymers and their colloidal assemblies using non-interfering techniques.

Understanding the microstructural parameters of amphiphilic copolymers that control the formation and structure of aggregated colloids (e.g., micelles) is essential for the rational design of hierarchically structured systems for applications in nanomedicine, personal care and food formulations. Although many analytical techniques have been employed to study such systems, in this investigation we adopted an integrated approach using non-interfering techniques - diffusion nuclear magnetic resonance (NMR) spectroscopy, dynamic light scattering (DLS) and synchrotron small-angle X-ray scattering (SAXS) - to probe the relationship between the microstructure of poly(ethylene glycol-b-caprolactone) (PEG-b-PCL) copolymers [e.g., block molecular weight (MW) and the mass fraction of PCL (fPCL)] and the structure of their aggregates. Systematic trends in the self-assembly behaviour were determined using a large family of well-defined block copolymers with variable PEG and PCL block lengths (number-average molecular weights (Mn) between 2 and 10 and 0.5-15 kDa, respectively) and narrow dispersity (Ð < 1.12). For all of the copolymers, a clear transition in the aggregate structure was observed when the hydrophobic fPCL was increased at a constant PEG block Mn, although the nature of this transition is also dependent on the PEG block Mn. Copolymers with low Mn PEG blocks (2 kDa) were observed to transition from unimers and loosely associated unimers to metastable aggregates and finally, to cylindrical micelles as the fPCL was increased. In comparison, copolymers with PEG block Mn of between 5 and 10 kDa transitioned from heterogenous metastable aggregates to cylindrical micelles and finally, well-defined ellipsoidal micelles (of decreasing aspect ratios) as the fPCL was increased. In all cases, the diffusion NMR spectroscopy, DLS and synchrotron SAXS results provided complementary information and the grounds for a phase diagram relating copolymer microstructure to aggregation behaviour and structure. Importantly, the absence of commonly depicted spherical micelles has implications for applications where properties may be governed by shape, such as, cellular uptake of nanomedicine formulations.

On the Efficacy of Water Transport in Leaves. A Coupled Xylem-Phloem Model of Water and Solute Transport.

In this paper, we present and use a coupled xylem/phloem mathematical model of passive water and solute transport through a reticulated vascular system of an angiosperm leaf. We evaluate the effect of leaf width-to-length proportion and orientation of second-order veins on the indexes of water transport into the leaves and sucrose transport from the leaves. We found that the most important factor affecting the steady-state pattern of hydraulic pressure distribution in the xylem and solute concentration in the phloem was leaf shape: narrower/longer leaves are less efficient in convecting xylem water and phloem solutes than wider/shorter leaves under all conditions studied. The degree of efficiency of transport is greatly influenced by the orientation of second-order veins relative to the main vein for all leaf proportions considered; the dependence is non-monotonic with efficiency maximized when the angle is approximately 45° to the main vein, although the angle of peak efficiency depends on other conditions. The sensitivity of transport efficiency to vein orientation increases with increasing vein conductivity. The vein angle at which efficiency is maximum tended to be smaller (relative to the main vein direction) in narrower leaves. The results may help to explain, or at least contribute to our understanding of, the evolution of parallel vein systems in monocot leaves.

Density functional theory of confined ionic liquids: the influence of power-law attractions on molecule distributions and surface forces.

Interaction energies and density profiles for two model ionic liquids, [C4mim+][BF4 -] and [C4mim+][TFSI-], confined between charged planar walls are studied within a density functional theory framework. The results of these simulations are also compared with results assuming a simpler linear hexamer-monomer, cation-anion system. We focus attention on the effect on the atom site distributions and the surface forces of an additional, specific attractive potential between oppositely charged molecules. We consider both short- and long-ranged attractive potentials in order to span the degree to which the ionic counterions associate. Independent of its strength, we interpret the results found with the short-ranged potential to be a manifestation of limited molecular association. In contrast, depending on its strength, the results found with the long-ranged potential suggest a much stronger and possibly longer ranged associations of ionic groups. Both potentials are found to influence the behavior of the surface force at small separations, while the long-ranged attractive potential has the greater influence of the two on the long-ranged behavior of the surface force.

A Comprehensive Biophysical Model of Ion and Water Transport in Plant Roots. III. Quantifying the Energy Costs of Ion Transport in Salt-Stressed Roots of Arabidopsis.

Salt stress defense mechanisms in plant roots, such as active Na+ efflux and storage, require energy in the form of ATP. Understanding the energy required for these transport mechanisms is an important step toward achieving an understanding of salt tolerance. However, accurate measurements of the fluxes required to estimate these energy costs are difficult to achieve by experimental means. As a result, the magnitude of the energy costs of ion transport in salt-stressed roots relative to the available energy is unclear, as are the relative contributions of different defense mechanisms to the total cost. We used mathematical modeling to address three key questions about the energy costs of ion transport in salt-stressed Arabidopsis roots: are the energy requirements calculated on the basis of flux data feasible; which transport steps are the main contributors to the total energy costs; and which transport processes could be altered to minimize the total energy costs? Using our biophysical model of ion and water transport we calculated the energy expended in the trans-plasma membrane and trans-tonoplast transport of Na+, K+, Cl-, and H+ in different regions of a salt-stressed model Arabidopsis root. Our calculated energy costs exceeded experimental estimates of the energy supplied by root respiration for high external NaCl concentrations. We found that Na+ exclusion from, and Cl- uptake into, the outer root were the major contributors to the total energy expended. Reducing the leakage of Na+ and the active uptake of Cl- across outer root plasma membranes would lower energy costs while enhancing exclusion of these ions. The high energy cost of ion transport in roots demonstrates that the energetic consequences of altering ion transport processes should be considered when attempting to improve salt tolerance.

Energy costs of salt tolerance in crop plants.

Agriculture is expanding into regions that are affected by salinity. This review considers the energetic costs of salinity tolerance in crop plants and provides a framework for a quantitative assessment of costs. Different sources of energy, and modifications of root system architecture that would maximize water vs ion uptake are addressed. Energy requirements for transport of salt (NaCl) to leaf vacuoles for osmotic adjustment could be small if there are no substantial leaks back across plasma membrane and tonoplast in root and leaf. The coupling ratio of the H+ -ATPase also is a critical component. One proposed leak, that of Na+ influx across the plasma membrane through certain aquaporin channels, might be coupled to water flow, thus conserving energy. For the tonoplast, control of two types of cation channels is required for energy efficiency. Transporters controlling the Na+ and Cl- concentrations in mitochondria and chloroplasts are largely unknown and could be a major energy cost. The complexity of the system will require a sophisticated modelling approach to identify critical transporters, apoplastic barriers and root structures. This modelling approach will inform experimentation and allow a quantitative assessment of the energy costs of NaCl tolerance to guide breeding and engineering of molecular components.

Statistical analysis and modeling of the geometry and topology of plant roots.

The root is an important organ of a plant since it is responsible for water and nutrient uptake. Analyzing and modeling variabilities in the geometry and topology of roots can help in assessing the plant's health, understanding its growth patterns, and modeling relations between plant species and between plants and their environment. In this article, we develop a framework for the statistical analysis and modeling of the geometry and topology of plant roots. We represent root structures as points in a tree-shape space equipped with a metric that quantifies geometric and topological differences between pairs of roots. We then use these building blocks to compute geodesics, i.e., optimal deformations under the metric between root structures, and to perform statistical analysis on root populations. We demonstrate the utility of the proposed framework through an application to a dataset of wheat roots grown in different environmental conditions. We also show that the framework can be used in various applications including classification and regression.

A Comprehensive Biophysical Model of Ion and Water Transport in Plant Roots. II. Clarifying the Roles of SOS1 in the Salt-Stress Response in Arabidopsis.

SOS1 transporters play an essential role in plant salt tolerance. Although SOS1 is known to encode a plasma membrane Na+/H+ antiporter, the transport mechanisms by which these transporters contribute to salt tolerance at the level of the whole root are unclear. Gene expression and flux measurements have provided conflicting evidence for the location of SOS1 transporter activity, making it difficult to determine their function. Whether SOS1 transporters load or unload Na+ from the root xylem transpiration stream is also disputed. To address these areas of contention, we applied a mathematical model to answer the question: what is the function of SOS1 transporters in salt-stressed Arabidopsis roots? We used our biophysical model of ion and water transport in a salt-stressed root to simulate a wide range of SOS1 transporter locations in a model Arabidopsis root, providing a level of detail that cannot currently be achieved by experimentation. We compared our simulations with available experimental data to find reasonable parameters for the model and to determine likely locations of SOS1 transporter activity. We found that SOS1 transporters are likely to be operating in at least one tissue of the outer mature root, in the mature stele, and in the epidermis of the root apex. SOS1 transporter activity in the mature outer root cells is essential to maintain low cytosolic Na+ levels in the root and also restricts the uptake of Na+ to the shoot. SOS1 transporters in the stele actively load Na+ into the xylem transpiration stream, enhancing the transport of Na+ and water to the shoot. SOS1 transporters acting in the apex restrict cytosolic Na+ concentrations in the apex but are unable to maintain low cytosolic Na+ levels in the mature root. Our findings suggest that targeted, tissue-specific overexpression or knockout of SOS1 may lead to greater salt tolerance than has been achieved with constitutive gene changes. Tissue-specific changes to the expression of SOS1 could be used to identify the appropriate balance between limiting Na+ uptake to the shoot while maintaining water uptake, potentially leading to enhancements in salt tolerance.

An Automatic Field Plot Extraction Method From Aerial Orthomosaic Images.

Unmanned aerial vehicles have an immense capacity for remote imaging of plants in agronomic field research trials. Traits extracted from the plots can explain development of the plants coverage, growth, flowering status, and related phenomenon. An important prerequisite step to obtain such information is to find the exact position of plots to extract them from an orthomosaic image. Extraction of plots using tools which assume a uniform spacing is often erroneous because the plots may neither be perfectly aligned nor equally distributed in a field. A novel approach is proposed which uses image-based optimization algorithm to find the alignment of plots. The method begins with a uniformly spaced grid of plots which is iteratively aligned with regions of high vegetation index, i.e., the underlying plots. The approach is validated and tested on two different orthomosaic images of fields containing wheat plots with simulated and real alignment problems, respectively. The result of alignment is compared to manually located ground truth position of plots and the errors are quantitatively analyzed. The effectiveness of the proposed method is confirmed in accurately estimating the phenotypic trait of canopy coverage compared to the common methods of extraction from uniform grids or trimmed grids. The software developed in this study is available from SourceForge, https://sourceforge.net/projects/phenalysis/.

Density functional theory of confined ionic liquids: A survey of the effects of ion type, molecular charge distribution, and surface adsorption.

Ionic liquids confined between two planar charged walls are explored using density functional theory. The effort represents a study of the effects of the molecular structure, molecular charge distribution, and degree of surface adsorption on forces between the surfaces and on the inhomogeneous atom density profiles. Surface adsorption was found to significantly affect both the magnitude and sign of the surface forces, while differences in the distribution of molecular charge did not. On the other hand, different bulk densities were found to produce dramatically different surface forces indicating a difference in the degree of molecular packing at and near surfaces. No long-range forces were found in any of the cases considered. We conclude that in the absence of any specific cation-anion pairing, surface charges are effectively screened, and the surface forces are dominated by short ranged steric and dispersion interactions between adsorbed molecular layers. In many cases, very similar surface forces correspond to very different molecular arrangements, suggesting that unambiguous interpretation of measured surface forces in ionic liquids, in terms of molecular behavior, may be difficult to guarantee.

High-Throughput Field Imaging and Basic Image Analysis in a Wheat Breeding Programme.

Visual assessment of colour-based traits plays a key role within field-crop breeding programmes, though the process is subjective and time-consuming. Digital image analysis has previously been investigated as an objective alternative to visual assessment for a limited number of traits, showing suitability and slight improvement to throughput over visual assessment. However, easily adoptable, field-based high-throughput methods are still lacking. The aim of the current study was to produce a high-throughput digital imaging and analysis pipeline for the assessment of colour-based traits within a wheat breeding programme. This was achieved through the steps of (i) a proof-of-concept study demonstrating basic image analysis methods in a greenhouse, (ii) application of these methods to field trials using hand-held imaging, and (iii) developing a field-based high-throughput imaging infrastructure for data collection. The proof of concept study showed a strong correlation (r = 0.95) between visual and digital assessments of wheat physiological yellowing (PY) in a greenhouse environment, with both scores having similar heritability (H2 = 0.85 and 0.76, respectively). Digital assessment of hand-held field images showed strong correlations to visual scores for PY (r = 0.61 and 0.78), senescence (r = 0.74 and 0.75) and Septoria tritici blotch (STB; r = 0.76), with greater heritability of digital scores, excluding STB. Development of the high-throughput imaging infrastructure allowed for images of field plots to be collected at a rate of 7,400 plots per hour. Images of an advanced breeding trial collected with this system were analysed for canopy cover at two time-points, with digital scores correlating strongly to visual scores (r = 0.88 and 0.86) and having similar or greater heritability. This study details how high-throughput digital phenotyping can be applied to colour-based traits within field trials of a wheat breeding programme. It discusses the logistics of implementing such systems with minimal disruption to the programme, provides a detailed methodology for the basic image analysis methods utilized, and has potential for application to other field-crop breeding or research programmes.

Correction to: Detection and analysis of wheat spikes using Convolutional Neural Networks.

[This corrects the article DOI: 10.1186/s13007-018-0366-8.].

Detection and analysis of wheat spikes using Convolutional Neural Networks.

BACKGROUND: Field phenotyping by remote sensing has received increased interest in recent years with the possibility of achieving high-throughput analysis of crop fields. Along with the various technological developments, the application of machine learning methods for image analysis has enhanced the potential for quantitative assessment of a multitude of crop traits. For wheat breeding purposes, assessing the production of wheat spikes, as the grain-bearing organ, is a useful proxy measure of grain production. Thus, being able to detect and characterize spikes from images of wheat fields is an essential component in a wheat breeding pipeline for the selection of high yielding varieties. RESULTS: We have applied a deep learning approach to accurately detect, count and analyze wheat spikes for yield estimation. We have tested the approach on a set of images of wheat field trial comprising 10 varieties subjected to three fertilizer treatments. The images have been captured over one season, using high definition RGB cameras mounted on a land-based imaging platform, and viewing the wheat plots from an oblique angle. A subset of in-field images has been accurately labeled by manually annotating all the spike regions. This annotated dataset, called SPIKE, is then used to train four region-based Convolutional Neural Networks (R-CNN) which take, as input, images of wheat plots, and accurately detect and count spike regions in each plot. The CNNs also output the spike density and a classification probability for each plot. Using the same R-CNN architecture, four different models were generated based on four different datasets of training and testing images captured at various growth stages. Despite the challenging field imaging conditions, e.g., variable illumination conditions, high spike occlusion, and complex background, the four R-CNN models achieve an average detection accuracy ranging from 88 to 94 % across different sets of test images. The most robust R-CNN model, which achieved the highest accuracy, is then selected to study the variation in spike production over 10 wheat varieties and three treatments. The SPIKE dataset and the trained CNN are the main contributions of this paper. CONCLUSION: With the availability of good training datasets such us the SPIKE dataset proposed in this article, deep learning techniques can achieve high accuracy in detecting and counting spikes from complex wheat field images. The proposed robust R-CNN model, which has been trained on spike images captured during different growth stages, is optimized for application to a wider variety of field scenarios. It accurately quantifies the differences in yield produced by the 10 varieties we have studied, and their respective responses to fertilizer treatment. We have also observed that the other R-CNN models exhibit more specialized performances. The data set and the R-CNN model, which we make publicly available, have the potential to greatly benefit plant breeders by facilitating the high throughput selection of high yielding varieties.

Land-based crop phenotyping by image analysis: consistent canopy characterization from inconsistent field illumination.

BACKGROUND: One of the main challenges associated with image-based field phenotyping is the variability of illumination. During a single day's imaging session, or between different sessions on different days, the sun moves in and out of cloud cover and has varying intensity. How is one to know from consecutive images alone if a plant has become darker over time, or if the weather conditions have simply changed from clear to overcast? This is a significant problem to address as colour is an important phenotypic trait that can be measured automatically from images. RESULTS: In this work we use an industry standard colour checker to balance the colour in images within and across every day of a field trial conducted over four months in 2016. By ensuring that the colour checker is present in every image we are afforded a 'ground truth' to correct for varying illumination conditions across images. We employ a least squares approach to fit a quadratic model for correcting RGB values of an image in such a way that the observed values of the colour checker tiles align with their true values after the transformation. CONCLUSIONS: The proposed method is successful in reducing the error between observed and reference colour chart values in all images. Furthermore, the standard deviation of mean canopy colour across multiple days is reduced significantly after colour correction is applied. Finally, we use a number of examples to demonstrate the usefulness of accurate colour measurements in recording phenotypic traits and analysing variation among varieties and treatments.

Pre-processing by data augmentation for improved ellipse fitting.

Ellipse fitting is a highly researched and mature topic. Surprisingly, however, no existing method has thus far considered the data point eccentricity in its ellipse fitting procedure. Here, we introduce the concept of eccentricity of a data point, in analogy with the idea of ellipse eccentricity. We then show empirically that, irrespective of ellipse fitting method used, the root mean square error (RMSE) of a fit increases with the eccentricity of the data point set. The main contribution of the paper is based on the hypothesis that if the data point set were pre-processed to strategically add additional data points in regions of high eccentricity, then the quality of a fit could be improved. Conditional validity of this hypothesis is demonstrated mathematically using a model scenario. Based on this confirmation we propose an algorithm that pre-processes the data so that data points with high eccentricity are replicated. The improvement of ellipse fitting is then demonstrated empirically in real-world application of 3D reconstruction of a plant root system for phenotypic analysis. The degree of improvement for different underlying ellipse fitting methods as a function of data noise level is also analysed. We show that almost every method tested, irrespective of whether it minimizes algebraic error or geometric error, shows improvement in the fit following data augmentation using the proposed pre-processing algorithm.

Land-based crop phenotyping by image analysis: Accurate estimation of canopy height distributions using stereo images.

In this paper we report on an automated procedure to capture and characterize the detailed structure of a crop canopy by means of stereo imaging. We focus attention specifically on the detailed characteristic of canopy height distribution-canopy shoot area as a function of height-which can provide an elaborate picture of canopy growth and health under a given set of conditions. We apply the method to a wheat field trial involving ten Australian wheat varieties that were subjected to two different fertilizer treatments. A novel camera self-calibration approach is proposed which allows the determination of quantitative plant canopy height data (as well as other valuable phenotypic information) by stereo matching. Utilizing the canopy height distribution to provide a measure of canopy height, the results compare favourably with manual measurements of canopy height (resulting in an R2 value of 0.92), and are indeed shown to be more consistent. By comparing canopy height distributions of different varieties and different treatments, the methodology shows that different varieties subjected to the same treatment, and the same variety subjected to different treatments can respond in much more distinctive and quantifiable ways within their respective canopies than can be captured by a simple trait measure such as overall canopy height.

Estimation of vegetation indices for high-throughput phenotyping of wheat using aerial imaging.

BACKGROUND: Unmanned aerial vehicles offer the opportunity for precision agriculture to efficiently monitor agricultural land. A vegetation index (VI) derived from an aerially observed multispectral image (MSI) can quantify crop health, moisture and nutrient content. However, due to the high cost of multispectral sensors, alternate, low-cost solutions have lately received great interest. We present a novel method for model-based estimation of a VI using RGB color images. The non-linear spatio-spectral relationship between the RGB image of vegetation and the index computed by its corresponding MSI is learned through deep neural networks. The learned models can be used to estimate VI of a crop segment. RESULTS: Analysis of images obtained in wheat breeding trials show that the aerially observed VI was highly correlated with ground-measured VI. In addition, VI estimates based on RGB images were highly correlated with VI deduced from MSIs. Spatial, spectral and temporal information of images contributed to estimation of VI. Both intra-variety and inter-variety differences were preserved by estimated VI. However, VI estimates were reliable until just before significant appearance of senescence. CONCLUSION: The proposed approach validates that it is reasonable to accurately estimate VI using deep neural networks. The results prove that RGB images contain sufficient information for VI estimation. It demonstrates that low-cost VI measurement is possible with standard RGB cameras.