High-throughput Field Phenotyping Research Articles

High-throughput field phenotyping in vineyards: demand, approaches, objectives

High-throughput field phenotyping in vineyards: demand, approaches, objectives

Erratum to: Mixing things up! Identifying early diversity benefits and facilitating the development of improved variety mixtures with high throughput field phenotyping

Erratum to: Mixing things up! Identifying early diversity benefits and facilitating the development of improved variety mixtures with high throughput field phenotyping

Multi temporal multispectral UAV remote sensing allows for yield assessment across European wheat varieties already before flowering.

High throughput field phenotyping techniques employing multispectral cameras allow extracting a variety of variables and features to predict yield and yield related traits, but little is known about which types of multispectral features are optimal to forecast yield potential in the early growth phase. In this study, we aim to identify multispectral features that are able to accurately predict yield and aid in variety classification at different growth stages throughout the season. Furthermore, we hypothesize that texture features (TFs) are more suitable for variety classification than for yield prediction. Throughout 2021 and 2022, a trial involving 19 and 18 European wheat varieties, respectively, was conducted. Multispectral images, encompassing visible, Red-edge, and near-infrared (NIR) bands, were captured at 19 and 22 time points from tillering to harvest using an unmanned aerial vehicle (UAV) in the first and second year of trial. Subsequently, orthomosaic images were generated, and various features were extracted, including single-band reflectances, vegetation indices (VI), and TFs derived from a gray level correlation matrix (GLCM). The performance of these features in predicting yield and classifying varieties at different growth stages was assessed using random forest models. Measurements during the flowering stage demonstrated superior performance for most features. Specifically, Red reflectance achieved a root mean square error (RMSE) of 52.4g m-2 in the first year and 64.4g m-2 in the second year. The NDRE VI yielded the most accurate predictions with an RMSE of 49.1g m-2 and 60.6g m-2, respectively. Moreover, TFs such as CONTRAST and DISSIMILARITY displayed the best performance in predicting yield, with RMSE values of 55.5g m-2 and 66.3g m-2 across the two years of trial. Combining data from different dates enhanced yield prediction and stabilized predictions across dates. TFs exhibited high accuracy in classifying low and high-yielding varieties. The CORRELATION feature achieved an accuracy of 88% in the first year, while the HOMOGENEITY feature reached 92% accuracy in the second year. This study confirms the hypothesis that TFs are more suitable for variety classification than for yield prediction. The results underscore the potential of TFs derived from multispectral images in early yield prediction and varietal classification, offering insights for HTP and precision agriculture alike.

High-throughput field phenotyping reveals that selection in breeding has affected the phenology and temperature response of wheat in the stem elongation phase.

Crop growth and phenology are driven by seasonal changes in environmental variables, with temperature as one important factor. However, knowledge about genotype-specific temperature response and its influence on phenology is limited. Such information is fundamental to improve crop models and adapt selection strategies. We measured the increase in height of 352 European winter wheat varieties in 4 years to quantify phenology, and fitted an asymptotic temperature response model. The model used hourly fluctuations in temperature to parameterize the base temperature (Tmin), the temperature optimum (rmax), and the steepness (lrc) of growth responses. Our results show that higher Tmin and lrc relate to an earlier start and end of stem elongation. A higher rmax relates to an increased final height. Both final height and rmax decreased for varieties originating from the continental east of Europe towards the maritime west. A genome-wide association study (GWAS) indicated a quantitative inheritance and a large degree of independence among loci. Nevertheless, genomic prediction accuracies (GBLUPs) for Tmin and lrc were low (r≤0.32) compared with other traits (r≥0.59). As well as known, major genes related to vernalization, photoperiod, or dwarfing, the GWAS indicated additional, as yet unknown loci that dominate the temperature response.

CottonSense: A high-throughput field phenotyping system for cotton fruit segmentation and enumeration on edge devices

High-throughput phenotyping (HTP) has become a powerful tool for gaining insights into the genetic and environmental factors that affect cotton (Gossypium spp.) growth and yield. With the recent advances in the field of computer vision, namely the integration of deep learning algorithms, the accuracy and efficiency of HTP systems have improved dramatically, enabling them to automatically quantify such fundamental phenotypic traits as fruit identification and enumeration. However, there is currently no HTP system available for counting all the reproductive phases of cotton crop that can be deployed in agronomic field conditions throughout the growing season. This study presents CottonSense, an advanced HTP system that overcomes the challenges of deployment across multiple growth periods by effectively segmenting and enumerating cotton fruits at four stages of growth, including square, flower, closed boll, and open boll. Consequently, CottonSense enhances agronomic management through increased opportunities for data collection and analysis. Using RGB-D cameras, it captures and processes both two and three-dimensional data, facilitating a wider range of phenotypic trait extractions such as crop biomass and plant architecture. To segment the cotton fruits, a Mask-RCNN model is trained and optimized for faster inference using TensorRT. The model yields an average AP score of 79% in segmentation across the four fruit categories. Moreover, the model's accuracy in estimating total fruit count per image is validated by a strong agreement with the counts given by ten domain experts, as reflected by an R2 value of 0.94. Furthermore, to accurately count the segmented fruits over large populations of plants, an enumeration algorithm based on a tracking strategy is developed that achieves an R2 value of 0.93 when compared to hand-counted fruits in the field. The proposed HTP system, which is implemented entirely on an edge computing device, is cost-effective and power-efficient, making it an effective tool for high-yield cotton breeding and crop improvement. The code for CottonSense is publicly available at https://github.com/FeriBolour/CottonSense.

Mixing things up! Identifying early diversity benefits and facilitating the development of improved variety mixtures with high throughput field phenotyping

AbstractCrop diversification is a potential strategy to increase the stability and productivity of crops, while reducing pathogen pressures and pesticide requirements. Crop variety mixtures provide some of these diversification benefits and their cultivation is fully compatible with current mechanized agronomic practices. However, the development of optimal variety mixtures is a long, labour‐intense process requiring extensive field trials. High throughput field phenotyping (HTFP) methods provide promising applications in field testing because they allow for precise, repeatable, and rapid measurements of crop properties. Here, we evaluated the use of HTFP for developing high‐performing oat (Avena sativa) variety mixtures by testing its suitability to predict diversity yield benefits from repeated canopy measurements across the growing season. Analyzing 26 mixtures of five varieties, we found significant overyielding at harvest, that is, mixtures were on average more productive than expected based on component pure stands. This grain yield overyielding was well predicted from deviations between mixture and pure stand canopy cover estimations, derived from HTFP mid‐way through the growing season. This shows that (i) positive interactions between oat varieties occur already at an early stage, (ii) such interactions lead to increased potential for light interception, (iii) HTFP offers rapid, scalable methods to screen for performant variety mixtures.

Image-based phenomic prediction can provide valuable decision support in wheat breeding

Key messageGenotype-by-environment interactions of secondary traits based on high-throughput field phenotyping are less complex than those of target traits, allowing for a phenomic selection in unreplicated early generation trials.Traditionally, breeders’ selection decisions in early generations are largely based on visual observations in the field. With the advent of affordable genome sequencing and high-throughput phenotyping technologies, enhancing breeders’ ratings with such information became attractive. In this research, it is hypothesized that GtimesE interactions of secondary traits (i.e., growth dynamics’ traits) are less complex than those of related target traits (e.g., yield). Thus, phenomic selection (PS) may allow selecting for genotypes with beneficial response-pattern in a defined population of environments. A set of 45 winter wheat varieties was grown at 5 year-sites and analyzed with linear and factor-analytic (FA) mixed models to estimate GtimesE interactions of secondary and target traits. The dynamic development of drone-derived plant height, leaf area and tiller density estimations was used to estimate the timing of key stages, quantities at defined time points and temperature dose–response curve parameters. Most of these secondary traits and grain protein content showed little GtimesE interactions. In contrast, the modeling of GtimesE for yield required a FA model with two factors. A trained PS model predicted overall yield performance, yield stability and grain protein content with correlations of 0.43, 0.30 and 0.34. While these accuracies are modest and do not outperform well-trained GS models, PS additionally provided insights into the physiological basis of target traits. An ideotype was identified that potentially avoids the negative pleiotropic effects between yield and protein content.

Phenotyping wood properties of Corymbia torelliana x Corymbia citriodora and Eucalyptus dunnii based on NIR spectra

ABSTRACT Selecting individuals in breeding programs for pulp production, the physical and chemical wood traits must be considered in the genotype selection. In this study, trees from hybrid progeny tests of Corymbia torelliana x Corymbia citriodora (CTOxCCT) and progenies of Eucalyptus dunnii (EDU) were investigated. Sawdust of genetically selected standing trees was collected, classified and prepared for near infrared (NIR) spectra readings. The chemical properties of the selected trees were determined through kraft pulping. Predictive models for each property were developed based on the reference data and NIR spectra. Two approaches for models were developed. The first approach, models were fitted with 25 samples of CTOxCCT, and in the second approach, models were fitted using 61 samples (25 of CTOxCCT and 36 of EDU). The estimated R²cv values were 0.60 and 0.73 for basic chip density, 0.37 and 0.65 for extractive contents, 0.56 and 0.53 for total lignin contents, 0.63 and 0.66 for S/G ratio, and 0.52 and 0.77 for screened pulp yield for the first and second approaches, respectively. All developed models have potential for ranking trees in breeding programs. NIR spectroscopy can potentially be applied as a high-throughput field phenotyping tool where thousands of varieties need to be evaluated.

Phenomic and genomic prediction of yield on multiple locations in winter wheat.

Genomic selection has recently become an established part of breeding strategies in cereals. However, a limitation of linear genomic prediction models for complex traits such as yield is that these are unable to accommodate Genotype by Environment effects, which are commonly observed over trials on multiple locations. In this study, we investigated how this environmental variation can be captured by the collection of a large number of phenomic markers using high-throughput field phenotyping and whether it can increase GS prediction accuracy. For this purpose, 44 winter wheat (Triticum aestivum L.) elite populations, comprising 2,994 lines, were grown on two sites over 2years, to approximate the size of trials in a practical breeding programme. At various growth stages, remote sensing data from multi- and hyperspectral cameras, as well as traditional ground-based visual crop assessment scores, were collected with approximately 100 different data variables collected per plot. The predictive power for grain yield was tested for the various data types, with or without genome-wide marker data sets. Models using phenomic traits alone had a greater predictive value (R2 = 0.39-0.47) than genomic data (approximately R2 = 0.1). The average improvement in predictive power by combining trait and marker data was 6%-12% over the best phenomic-only model, and performed best when data from one full location was used to predict the yield on an entire second location. The results suggest that genetic gain in breeding programmes can be increased by utilisation of large numbers of phenotypic variables using remote sensing in field trials, although at what stage of the breeding cycle phenomic selection could be most profitably applied remains to be answered.

A Low-Cost Sensorized Vehicle for In-Field Crop Phenotyping

The development of high-throughput field phenotyping, which uses modern detection technologies and advanced data processing algorithms, could increase productivity and make in-field phenotypic evaluation more efficient by collecting large amounts of data with no or minimal human assistance. Moreover, high-throughput plant phenotyping systems are also very effective in selecting crops and characterizing germplasm for drought tolerance and disease resistance by using spectral sensor data in combination with machine learning. In this study, an affordable high-throughput phenotyping platform (phenomobile) aims to obtain solutions at reasonable prices for all the components that make up it and the many data collected. The goal of the practical innovation in field phenotyping is to implement high-performance precision phenotyping under real-world conditions at accessible costs, making real-time data analysis techniques more user-friendly. This work aims to test the ability of a phenotyping prototype system constituted by an electric phenomobile integrated with a MAIA multispectral camera for real in-field plant characterization. This was done by acquiring spectral signatures of F1 hybrid Elisir (Olter Sementi) tomato plants and calculating their vegetation indexes. This work allowed to collect, in real time, a great number of field data about, for example, the morphological traits of crops, plant physiological activities, plant diseases, fruit maturity, and plant water stress.

Design of a Lightweight Autonomous Robotic Phenotyping System for Turfgrass

High-throughput field inspection and phenotyping are among the most promising applications for agricultural robotics. We present an autonomous, lightweight robotic system for turfgrass phenotyping applications. The robot is built on two integrated bikebot platform with high-clearance and low soil compaction. To improve the motion control accuracy, an invariant extended Kalman filter is built with fusion of GPS and inertial measurement unit for outdoor localization and navigation. Multiple cameras are used for multi-modal imaging collections. Preliminary experiment results show that the localization accuracy is at the centimeter level and the path-following error is less than 0.5 m most time for various reference trajectories in the feld. The robot design and autonomous navigation demonstrates the feasibility for turfgrass inspection application.

Breeding and Selection of Nursery Plants Assisted by High-Throughput Field Phenotyping Using UAV Imagery: Case Studies with Sweet Box (Sarcococca) and Garden Rose (Rosa)

Breeding and selection of nursery plants require evaluation of a wide variety of traits. Characteristics that are visually scored in the field include aesthetic beauty as well as tolerance to (a)biotic stresses, among others. This work proposes methodologies based on vegetation indices and canopy height data derived from visual imagery captured using an RGB (red, green, blue) camera embedded in a drone to rank and select genotypes. This approach relies on quantitative evaluation standards that exclude breeder bias and speed up data collection. A proof of concept for nursery plants was developed in two woody ornamentals: sweet box (Sarcococca Lindl.) and garden rose (Rosa L.). This work aimed to compare methodologies and to propose how drones (unmanned aerial vehicles, UAV) and high throughput field phenotyping (HTFP) can be used in a commercial plant selection program. Data was gathered in 2019 in three well-established breeding trials, two for sweet box and one for rose. Characteristics discussed include plant architecture traits such as plant height and shape in the sweet box and rose as well as floribundity, continuous flowering and disease resistance in roses. Correlations were calculated between on-ground measurements and UAV-derived methods, obtaining significant results. The advantages and disadvantages of the methodology and the approach for each of the traits are discussed.

Applications of Artificial Intelligence in Climate-Resilient Smart-Crop Breeding

Recently, Artificial intelligence (AI) has emerged as a revolutionary field, providing a great opportunity in shaping modern crop breeding, and is extensively used indoors for plant science. Advances in crop phenomics, enviromics, together with the other “omics” approaches are paving ways for elucidating the detailed complex biological mechanisms that motivate crop functions in response to environmental trepidations. These “omics” approaches have provided plant researchers with precise tools to evaluate the important agronomic traits for larger-sized germplasm at a reduced time interval in the early growth stages. However, the big data and the complex relationships within impede the understanding of the complex mechanisms behind genes driving the agronomic-trait formations. AI brings huge computational power and many new tools and strategies for future breeding. The present review will encompass how applications of AI technology, utilized for current breeding practice, assist to solve the problem in high-throughput phenotyping and gene functional analysis, and how advances in AI technologies bring new opportunities for future breeding, to make envirotyping data widely utilized in breeding. Furthermore, in the current breeding methods, linking genotype to phenotype remains a massive challenge and impedes the optimal application of high-throughput field phenotyping, genomics, and enviromics. In this review, we elaborate on how AI will be the preferred tool to increase the accuracy in high-throughput crop phenotyping, genotyping, and envirotyping data; moreover, we explore the developing approaches and challenges for multiomics big computing data integration. Therefore, the integration of AI with “omics” tools can allow rapid gene identification and eventually accelerate crop-improvement programs.

Evaluating biostimulants via high-throughput field phenotyping: Biophysical traits retrieval through PROSAIL inversion

Drought management largely depends on the availability of timely, accurate and integrated information about its characteristics. Concurrently, biostimulants could represent a sustainable measure to foster the resilience of cropping systems under water-limited conditions. Nevertheless, scientific recognition of the potential of biostimulants has not grown as fast as the interest from industry: therefore, there is an urgent need to investigate biostimulant action. In recent decades, remote sensing has been successfully applied to crop growth and stress monitoring. The use of radiative transfer models, often rooted in artificial intelligence, to estimate plant traits from remotely sensed data can be considered the link between the generalisation and the spatialisation of data, bringing field phenotyping and remote sensing closer together. In this framework, the present study was designed as a factorial combination of irrigation treatment (3 levels) and biostimulant treatment (3 levels) and conducted on processing tomato in open field during the summer of 2020. PROSAIL inversion was carried out to retrieve three major biophysical traits (LAI, LCC, CCC). The parameters dynamics during the season were investigated through GAM modelling. The validation of the model was carried out with a positive outcome in terms of accuracy. The PROSAIL inversion enabled the efficient retrieval of LAI and LCC at rates comparable to those in literature, while performing worse than literature findings only for CCC, probably due to the characteristics of tomato canopy. At the same time, the effect of irrigation was detected both for yield and quality data and detected through the GAM modelisation of the parameters. However, no biostimulant effect could be detected. The internal variability per plot of the retrieved biophysical traits was high. This, jointly with the uncertainty surrounding biostimulant testing and the magnitude of biostimulant effects, corroborated by the absence of results regarding biostimulant effect on yield and DM, lead to hypothesise that the bottleneck was linked to the biostimulant effect itself.

Tree architecture, light interception and water-use related traits are controlled by different genomic regions in an apple tree core collection.

Summary Tree architecture shows large genotypic variability, but how this affects water‐deficit responses is poorly understood. To assess the possibility of reaching ideotypes with adequate combinations of architectural and functional traits in the face of climate change, we combined high‐throughput field phenotyping and genome‐wide association studies (GWAS) on an apple tree (Malus domestica) core‐collection.We used terrestrial light detection and ranging (T‐LiDAR) scanning and airborne multispectral and thermal imagery to monitor tree architecture, canopy shape, light interception, vegetation indices and transpiration on 241 apple cultivars submitted to progressive field soil drying. GWAS was performed with single nucleotide polymorphism (SNP)‐by‐SNP and multi‐SNP methods.Large phenotypic and genetic variability was observed for all traits examined within the collection, especially canopy surface temperature in both well‐watered and water deficit conditions, suggesting control of water loss was largely genotype‐dependent. Robust genomic associations revealed independent genetic control for the architectural and functional traits. Screening associated genomic regions revealed candidate genes involved in relevant pathways for each trait.We show that multiple allelic combinations exist for all studied traits within this collection. This opens promising avenues to jointly optimize tree architecture, light interception and water use in breeding strategies. Genotypes carrying favourable alleles depending on environmental scenarios and production objectives could thus be targeted.

Outdoor Plant Segmentation With Deep Learning for High-Throughput Field Phenotyping on a Diverse Wheat Dataset.

Robust and automated segmentation of leaves and other backgrounds is a core prerequisite of most approaches in high-throughput field phenotyping. So far, the possibilities of deep learning approaches for this purpose have not been explored adequately, partly due to a lack of publicly available, appropriate datasets. This study presents a workflow based on DeepLab v3+ and on a diverse annotated dataset of 190 RGB (350 x 350 pixels) images. Images of winter wheat plants of 76 different genotypes and developmental stages have been acquired throughout multiple years at high resolution in outdoor conditions using nadir view, encompassing a wide range of imaging conditions. Inconsistencies of human annotators in complex images have been quantified, and metadata information of camera settings has been included. The proposed approach achieves an intersection over union (IoU) of 0.77 and 0.90 for plants and soil, respectively. This outperforms the benchmarked machine learning methods which use Support Vector Classifier and/or Random Forrest. The results show that a small but carefully chosen and annotated set of images can provide a good basis for a powerful segmentation pipeline. Compared to earlier methods based on machine learning, the proposed method achieves better performance on the selected dataset in spite of using a deep learning approach with limited data. Increasing the amount of publicly available data with high human agreement on annotations and further development of deep neural network architectures will provide high potential for robust field-based plant segmentation in the near future. This, in turn, will be a cornerstone of data-driven improvement in crop breeding and agricultural practices of global benefit.

A Review of High-Throughput Field Phenotyping Systems: Focusing on Ground Robots.

Manual assessments of plant phenotypes in the field can be labor-intensive and inefficient. The high-throughput field phenotyping systems and in particular robotic systems play an important role to automate data collection and to measure novel and fine-scale phenotypic traits that were previously unattainable by humans. The main goal of this paper is to review the state-of-the-art of high-throughput field phenotyping systems with a focus on autonomous ground robotic systems. This paper first provides a brief review of nonautonomous ground phenotyping systems including tractors, manually pushed or motorized carts, gantries, and cable-driven systems. Then, a detailed review of autonomous ground phenotyping robots is provided with regard to the robot's main components, including mobile platforms, sensors, manipulators, computing units, and software. It also reviews the navigation algorithms and simulation tools developed for phenotyping robots and the applications of phenotyping robots in measuring plant phenotypic traits and collecting phenotyping datasets. At the end of the review, this paper discusses current major challenges and future research directions.

Development of a high-throughput field phenotyping rover optimized for size-limited breeding fields as open-source hardware.

Phenotyping is a critical process in plant breeding, especially when there is an increasing demand for streamlining a selection process in a breeding program. Since manual phenotyping has limited efficiency, high-throughput phenotyping methods are recently popularized owing to progress in sensor and image processing technologies. However, in a size-limited breeding field, which is common in Japan and other Asian countries, it is challenging to introduce large machinery in the field or fly unmanned aerial vehicles over the field. In this study, we developed a ground-based high-throughput field phenotyping rover that could be easily introduced to a field regardless of the scale and location of the field even without special facilities. We also made the field rover open-source hardware, making its system available to public for easy modification, so that anyone can build one for their own use at a low cost. The trial run of the field rover revealed that it allowed the collection of detailed remote-sensing images of plants and quantitative analyses based on the images. The results suggest that the field rover developed in this study could allow efficient phenotyping of plants especially in a small breeding field.

Phenomics data processing: A plot-level model for repeated measurements to extract the timing of key stages and quantities at defined time points

• General plot-level model for repeated high-throughput field phenotyping measurements. • Extraction of three main intermediate trait categories for dynamic modelling. • Seamless processing approach that integrates temporal and spatial modelling. • Phenomics data processing cheatsheet. Decision-making in breeding increasingly depends on the ability to capture and predict crop responses to changing environmental factors. Advances in crop modeling as well as high-throughput field phenotyping (HTFP) hold promise to provide such insights. Processing HTFP data is an interdisciplinary task that requires broad knowledge on experimental design, measurement techniques, feature extraction, dynamic trait modeling, and prediction of genotypic values using statistical models. To get an overview of sources of variation in HTFP, we develop a general plot-level model for repeated measurements. Based on this model, we propose a seamless step-wise procedure that allows for carry on of estimated means and variances from stage to stage. The process builds on the extraction of three intermediate trait categories; (1) timing of key stages, (2) quantities at defined time points or periods, and (3) dose-response curves. In a first stage, these intermediate traits are extracted from low-level traits’ time series (e.g., canopy height) using P-splines and the quarter of maximum elongation rate method (QMER), as well as final height percentiles. In a second and third stage, extracted traits are further processed using a stage-wise linear mixed model analysis. Using a wheat canopy growth simulation to generate canopy height time series, we demonstrate the suitability of the stage-wise process for traits of the first two above-mentioned categories. Results indicate that, for the first stage, the P-spline/QMER method was more robust than the percentile method. In the subsequent two-stage linear mixed model processing, weighting the second and third stage with error variance estimates from the previous stages improved the root mean squared error. We conclude that processing phenomics data in stages represents a feasible approach if estimated means and variances are carried forward from one processing stage to the next. P-splines in combination with the QMER method are suitable tools to extract timing of key stages and quantities at defined time points from HTFP data.

High-throughput field phenotyping of soybean: Spotting an ideotype

Soybean is among the most important crops for food and feed production worldwide. Sustainable and local production in regions with marginal climates requires cold-adapted varieties that create high yield and protein content in a short vegetation period. Drone-based high-throughput field phenotyping methods allow monitoring the success and the developmental speed of genotypes in such target environments. This study exemplifies that such frequent and precise analyses of remotely sensed canopy growth traits can be used to derive the optimal genotype, a so-called ideotype, for a given mega-environment. For the case example of Switzerland, a country with a temperate oceanic climate, the results indicate that image-derived traits allow predicting yield and protein content from the dynamics of vegetative growth. Genotypes with early canopy cover produce high yield, whereas genotypes that show a prolonged duration until they have reached their final maximum of leaf area index are characterized by a high protein content. Analyses of early performance trial stage material indicate that there are genotypes that combine both features of growth dynamics. Whether these genotypes are then indeed successful in breeding programs remains to be investigated, since this also depends on disease resistance and other traits of those genotypes. Yet, overall, this study provides strong indications of the high value of high-throughput field phenotyping in the context of physiological and breeding-related analyses of crops.