Seed-per-pod estimation for plant breeding using deep learning

Uzal, Lucas C.; Grinblat, Guillermo L.; Namías, Rafael; Larese, Mónica G.; Bianchi, Julieta Sofia; Morandi, E. N.; Granitto, Pablo M.

doi:10.1016/j.compag.2018.04.024

Cited by 122 publications

(65 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A recent study by Zhang et al (2019) provides a procedure using computer vision for counting soybean pod under experimental settings that would address the phenotypic limitation of this study. Similarly, Uzal et al (2018) and Li et al (2019) recently proposed an imagery system for counting seeds directly from images of soybean pods, yet another yield component limited by the challenging phenotyping. Technologies that enable better, faster and cheaper data collection remain a key limiting factor for the research in yield components.…”

Section: Phenotypingmentioning

confidence: 99%

Quantitative Genomic Dissection of Soybean Yield Components

Xavier

Rainey

2020

G3 Genes|Genomes|Genetics

View full text Add to dashboard Cite

Soybean is a crop of major economic importance with low rates of genetic gains for grain yield compared to other field crops. A deeper understanding of the genetic architecture of yield components may enable better ways to tackle the breeding challenges. Key yield components include the total number of pods, nodes and the ratio pods per node. We evaluated the SoyNAM population, containing approximately 5600 lines from 40 biparental families that share a common parent, in 6 environments distributed across 3 years. The study indicates that the yield components under evaluation have low heritability, a reasonable amount of epistatic control, and partially oligogenic architecture: 18 quantitative trait loci were identified across the three yield components using multi-approach signal detection. Genetic correlation between yield and yield components was highly variable from family-to-family, ranging from -0.2 to 0.5. The genotype-by-environment correlation of yield components ranged from -0.1 to 0.4 within families. The number of pods can be utilized for indirect selection of yield. The selection of soybean for enhanced yield components can be successfully performed via genomic prediction, but the challenging data collections necessary to recalibrate models over time makes the introgression of QTL a potentially more feasible breeding strategy. The genomic prediction of yield components was relatively accurate across families, but less accurate predictions were obtained from within family predictions and predicting families not observed included in the calibration set.

show abstract

Section: Phenotypingmentioning

confidence: 99%

Quantitative Genomic Dissection of Soybean Yield Components

Xavier

Rainey

2020

G3 Genes|Genomes|Genetics

View full text Add to dashboard Cite

show abstract

“…As deep learning methods for computer vision become increasingly population for phenotyping morphological traits (Singh et al 2018), the current limitations with data collection could be addressed by an automated high-throughput phenotyping instead of human counts, that would likely increase both accuracy and scalability of the process. For instance, both Uzal et al (2018) and Li et al (2019) recently proposed an imagery system for counting seeds directly from images of soybean pods, yet another yield component limited by the challenging phenotyping. Technologies that enable better, faster and cheaper data collection remain a key limiting factor to research on yield components.…”

Section: Discussionmentioning

confidence: 99%

Quantitative Genomic Dissection of Soybean Yield Components

Xavier

Rainey

2019

Preprint

View full text Add to dashboard Cite

12Soybean is a crop of major economic importance with low rates of genetic gains for grain yield 13 compared to other field crops. A deeper understanding of the genetic architecture of yield 14 components may enable better ways to tackle the breeding challenges. Key yield components 15 include the total number of pods, nodes and the ratio pods per node. We evaluated the SoyNAM 16 population, containing approximately 5600 lines from 40 biparental families that share a common 17 parent, in 6 environments distributed across 3 years. The study indicates that the yield components 18 under evaluation have low heritability, a reasonable amount of epistatic control, and partially 19 oligogenic architecture: 18 quantitative trait loci were identified across the three yield components 20 using multi-approach signal detection. Genetic correlation between yield and yield components was 21 highly variable from family-to-family, ranging from -0.2 to 0.5. The genotype-by-environment 22 correlation of yield components ranged from -0.1 to 0.4 within families. The number of pods can be 23 utilized for indirect selection of yield. The selection of soybean for enhanced yield components can 24 be successfully performed via genomic prediction, but the challenging data collections necessary to 25 recalibrate models over time makes the introgression of QTLs a potentially more feasible breeding 26 strategy. The genomic prediction of yield components was relatively accurate across families, but 27 less accurate predictions were obtained from within-family predictions and predicting families not 28 observed included in the calibration set. 29 30 Introduction 31 32Soybean is a field crop of major importance due to its seed composition, containing approximately 33 40% protein and 20% oil. Soybean has restricted genetic basis (Mikel et al. 2010) and the rate of 34

show abstract

“…One characteristic of aerial photos is that they are usually taken for a well-defined purpose at identical heights, using identical camera settings. The purpose can be to analyse a specific agricultural land, to track down areas [5] contaminated by weeds, or to search for any other kind of changes.…”

Section: A Deep Learningmentioning

confidence: 99%

Automated evaluation of agricultural damage using UAV survey

Stojcsics

Domozi

Molnár

2018

Acta Universitatis Sapientiae, Agriculture and Environment

View full text Add to dashboard Cite

In the last decade, the rate of the industrial usage of fixed-wing and blended wing aircraft has increased. A 1–2-km2 area can be surveyed by such a drone within 30 to 60 minutes, without any special infrastructure, and this can be repeated at any time. This provides an opportunity to conduct automatized surveys and time series data testing, which can be used as a basis to decide specific processes. The state and the development of the plants can be monitored as well as the spread of pests and the efficiency of the procedures that protect against them. During the surveys, thousands of images are taken of the area, which can be converted to a georeferenced large-sized map within 20 to 40 hours, including post-production and a resolution varying from 0.01 to 0.1 cm/pixel. The paper provides a solution to the industrial post-production of these high-quantity data, in which a deep learning-based automated process using Matlab is presented, including a comparison of the results to the GIS data.

show abstract

Seed-per-pod estimation for plant breeding using deep learning

Cited by 122 publications

References 34 publications

Quantitative Genomic Dissection of Soybean Yield Components

Quantitative Genomic Dissection of Soybean Yield Components

Quantitative Genomic Dissection of Soybean Yield Components

Automated evaluation of agricultural damage using UAV survey

Contact Info

Product

Resources

About