Osvaldo Martín Pérez

Dr. Osvaldo Martín Pérez

Instituto Nacional de Investigación Agropecuaria (INIA), Colonia del Sacramento, Colonia, Uruguay
orcid

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 November 2020 in Remote Sensing
Verified
Journal Article
Open access
PDF file
Rodrigo TrevisanRodrigo Trevisan
Rodrigo Trevisan
OPOsvaldo Martín Pérez
Osvaldo Martín Pérez
NSNathan Schmitz
Nathan Schmitz
BDBrian Diers
Brian Diers
NMNicolas Martin
Nicolas Martin
Soybean maturity is a trait of critical importance for the development of new soybean cultivars, nevertheless, its characterization based on visual ratings has many challenges. Unmanned aerial vehicles (UAVs) imagery-based high-throughput phenotyping methodologies have been proposed as an alternative to the traditional visual ratings of pod senescence. However, the lack of scalable and accurate methods to extract the desired information from the images remains a significant bottleneck in breeding programs. The objective of this study was to develop an image-based high-throughput phenotyping system for evaluating soybean maturity in breeding programs. Images were acquired twice a week, starting when the earlier lines began maturation until the latest ones were mature. Two complementary convolutional neural networks (CNN) were developed to predict the maturity date. The first using a single date and the second using the five best image dates identified by the first model. The proposed CNN architecture was validated using more than 15,000 ground truth observations from five trials, including data from three growing seasons and two countries. The trained model showed good generalization capability with a root mean squared error lower than two days in four out of five trials. Four methods of estimating prediction uncertainty showed potential at identifying different sources of errors in the maturity date predictions. The architecture developed solves limitations of previous research and can be used at scale in commercial breeding programs.

3 April 2019 in Nutrient Cycling in Agroecosystems
Verified
Journal Article
Closed access
PDF file
ABAndrés Beretta-Blanco
Andrés Beretta-Blanco
OPOsvaldo Martín Pérez
Osvaldo Martín Pérez
LCLeonidas Carrasco-Letelier
Leonidas Carrasco-Letelier

21 November 2024 in Remote Sensing
Verified
Journal Article
Open access
PDF file
OPOsvaldo Martín Pérez
Osvaldo Martín Pérez
BDBrian Diers
Brian Diers
NMNicolas Martin
Nicolas Martin
Several studies have used aerial images to predict physiological maturity (R8 stage) in soybeans (Glycine max (L.) Merr.). However, information for making predictions in the current growing season using models fitted in previous years is still necessary. Using the Random Forest machine learning algorithm and time series of RGB (red, green, blue) and multispectral images taken from a drone, this work aimed to study, in three breeding experiments of plant rows, how maturity predictions are impacted by a number of factors. These include the type of camera used, the number and time between flights, and whether models fitted with data obtained in one or more environments can be used to make accurate predictions in an independent environment. Applying principal component analysis (PCA), it was found that compared to the full set of 8–10 flights (R2 = 0.91–0.94; RMSE = 1.8–1.3 days), using data from three to five fights before harvest had almost no effect on the prediction error (RMSE increase ~0.1 days). Similar prediction accuracy was achieved using either a multispectral or an affordable RGB camera, and the excess green index (ExG) was found to be the important feature in making predictions. Using a model trained with data from two previous years and using fielding notes from check cultivars planted in the test season, the R8 stage was predicted, in 2020, with an error of 2.1 days. Periodically adjusted models could help soybean breeding programs save time when characterizing the cycle length of thousands of plant rows each season.

30 December 2017 in Spanish Journal of Agricultural Research
Verified
Journal Article
Open access
Pablo Martin Gonzalez BarriosPablo Martin Gonzalez Barrios
Pablo Martin Gonzalez Barrios
MCMarina Castro
Marina Castro
OPOsvaldo Martín Pérez
Osvaldo Martín Pérez
DVDiego Vilaró
Diego Vilaró
LGLucia Gutierrez
Lucia Gutierrez
Modeling genotype by environment interaction (GEI) is one of the most challenging aspects of plant breeding programs. The use of efficient trial networks is an effective way to evaluate GEI to define selection strategies. Furthermore, the experimental design and the number of locations, replications, and years are crucial aspects of multi-environment trial (MET) network optimization. The objective of this study was to evaluate the efficiency and performance of a MET network of sunflower (Helianthus annuus L.). Specifically, we evaluated GEI in the network by delineating mega-environments, estimating genotypic stability and identifying relevant environmental covariates. Additionally, we optimized the network by comparing experimental design efficiencies. We used the National Evaluation Network of Sunflower Cultivars of Uruguay (NENSU) in a period of 20 years. MET plot yield and flowering time information was used to evaluate GEI. Additionally, meteorological information was studied for each sunflower physiological stage. An optimal network under these conditions should have three replications, two years of evaluation and at least three locations. The use of incomplete randomized block experimental design showed reasonable performance. Three mega-environments were defined, explained mainly by different management of sowing dates. Late sowings dates had the worst performance in grain yield and oil production, associated with higher temperatures before anthesis and fewer days allocated to grain filling. The optimization of MET networks through the analysis of the experimental design efficiency, the presence of GEI, and appropriate management strategies have a positive impact on the expression of yield potential and selection of superior cultivars.

1 January 2020 in Agrociencia Uruguay
Verified
Journal Article
Open access
PDF file
OPOsvaldo Martín Pérez
Osvaldo Martín Pérez
LVLuis Viega
Luis Viega
MCMarina Castro
Marina Castro
Drought stress during wheat (Triticum aestivum L.) grain filling can affect the grain yield and its quality; effects that in some regions can be increased by the event La Nina, the cold phase of the climatic phenomenon El Nino Southern Oscillation. The aim was to evaluate if the evolution of relative water content (RWC) and stomatal conductance (g(s)) can be related to grain yield and its quality. Four spring wheat cultivars grown in the southern cone of South America were subjected to two levels of irrigation applied after anthesis, a wellwatered treatment (Control), and another with 50 % of the Control (Stress). RWC and g(s) were determined weekly during the stress period and agronomic traits at harvest. With a mean grain yield decrease of 14.8 % (P<0.05), two types of response to water deficit were identified for RWC and gs. Two cultivars (Biointa 1001 and LE2333) had slower RWC and gs decreases, which were related to a lower ratio of kernels to aerial biomass (P<0.05), and to a non-significant effect in kernel weight. While, in the other cultivars (LE2249 and LE2331), the water deficit caused a more rapid RWC and g(s) decrease related to a higher ratio of kernels to aerial biomass, and a significant kernel weight decrease (10.4 and 20.7 %, respectively). The latter cultivars, that had more rapid leaf dehydration under water stress, may also have been source-limited during the grain-filling period.

21 August 2025 in Agrociencia Uruguay
Verified
Journal Article
Open access
PDF file
OPOsvaldo Martín Pérez
Osvaldo Martín Pérez
Plant breeding programs know the advantages of high-throughput phenotyping (HTP) in increasing efficiency over classical phenotyping and screening methods, which is achieved by saving time and improving selection accuracy. Even so, most programs have not yet systematically implemented this technology into their breeding pipelines. This review aims to indicate the restrictions of implementing HTP at a large scale and to summarize studies according to the used devices, data classes collected, and artificial intelligence (AI) methods applied to predict and classify agronomic traits in plant breeding programs with a focus on soybean [Glycine max (L.) Merr.]. Excluding HTP platforms in laboratories and greenhouses, satellite remote sensing, and autonomous mobile robots, this review focuses on field-based HTP platforms that take aerial images from drones and apply AI methods to associate those images with the traits of interest. Field-based HTP research is also conducted using hand-held devices that record individual vegetation indices (e.g., NDVI), a few spectral bands (multispectral radiometers), or the continuous range of the electromagnetic light spectrum (spectroradiometers). However, plant breeders must evaluate thousands of experimental lines each year, so using these devices instead of drones implies a trade-off between acquisition accuracy and the time it takes to collect the data. A challenge in the coming years is fine-tuning scalable, reliable models and optimizing data input, processing, and output pipelines to provide breeders with helpful information before they make selections.