Embedded Files
October 7 to 12, 2024
Montevideo, Uruguay
difusion_agrokhipux_post.mp4Description
Description
The South American region (Argentina, Brazil, and Uruguay) has traditionally been a producer and exporter of agri-food products. Currently, one of the challenges these countries face in the 21st century is the loss of competitiveness associated with the environmental impacts of agriculture, increasing demand, consumers' growing concerns about safety, and a higher frequency of extreme weather events favored by climate change. Moreover, the emergence of other producing regions capable of applying new technologies poses new challenges for the agriculture of our countries.
While the physical characteristics of the soil generate heterogeneous and consistent productivity zones over the years, specific climate and/or crop management during the growing cycle can alternatively decrease or exacerbate production variability within the plot. Therefore, understanding the underlying factors of soil heterogeneity at the plot level and how weather conditions can amplify their effects on plant development and productivity is crucial for crop management. This knowledge can be useful for producers to more efficiently manage inputs and for researchers to identify genotypes with desired traits evaluated under real field conditions.
Biotechnological-agronomic research teams with access to massive biological sequencing data and the capability to select or modify organisms based on their genomic sequences face the multidimensional complexity of gene expression. Currently, machine learning tools are being employed to address this multidimensionality, enabling the tackling of various problems, including crop analysis. Part of this information is provided through the capture and analysis of computer vision at different scales—satellite, proximal, or mounted on autonomous vehicles. Training in these two areas becomes essential, both for the correct and informed adoption of imported technologies and for the generation of indigenous and appropriate knowledge in our region.
In the context outlined above, this course is introduced at the local level as a first approach to this subject. To carry out this course, a group of researchers from three countries has been brought together, possessing expertise in both basic and applied research in the fields of biotechnology, artificial intelligence, computer vision, and the development of scientific equipment in open hardware mode. These researchers have formed an academic network, initially funded with extraregional resources, which is now poised to be further developed and expanded. This course, therefore, constitutes a significant contribution to achieving this new phase through the interactions established among the participating researchers and students.
An important component of this course involves the study of case applications of machine learning and computer vision in agriculture: various tools will be addressed through examples applied in agriculture and viticulture throughout the course. These tools allow for the characterization of spatial variability within plots, enabling, for instance, the optimization of resource management (water, fertilizers, pesticides, etc.). The heterogeneity of the climate, coupled with uneven topography and soil characteristics, results in a high spatial and temporal variability in plant growth, which can be measured through various computer vision or machine learning tools.
Summary of Topics
Summary of Topics
Foundations of Machine Learning.
A) Introduction to Machine Learning and its Use in Agriculture and Plant Biology
B) Machine Learning Models
C) Data Preprocessing
D) Model Evaluation
Case Studies of Machine Learning Application in Agriculture
A) Detection of Plant Diseases
B) Crop Yield Prediction
C) Optimization of Agricultural Resources
Fundamentals of Computer Vision
A) Introduction to Computer Vision
B) Object Detection and Tracking
C) Pattern Recognition
D) Presentation and Discussion of Computer Vision Applications in Different Crops (coffee, oranges, apples, grapes).
Fundamentals of Open Hardware Design for Scientific Equipment
A) Basic Concepts of Open Hardware
B) Design of Sensors and Devices for Plant Phenotyping
C) Development of Open Hardware Platforms
Multispectral Plant Analysis
A) Fundamentals of Spectroscopy
B) Multispectral and Hyperspectral Sensors
C) Acquisition and Processing of Multispectral Data
D) Feature Extraction and Data Analysis
E) Applications in Plant Phenotyping
F) Integration with Machine Learning and Computer Vision
G) Case Studies in Agriculture
Workshop on Assembling a Plant Phenotyping Device.
Hands-on Image Analysis Workshop
Hands-on Machine Learning Workshop
Discussion of Scientific Articles
Roundtable on the Impact of New Technologies in Agriculture
Open space for the discussion of technical and social aspects of implementing the techniques covered in the course into agricultural practice.
Visit to an experimental field and meeting with producers
The course includes opportunities for each student to explore ways to apply the principles covered in the course to their projects.
Format and Schedule
Format and Schedule
Format: In-person - Hours: 44 (18 theoretical, 26 practical). The course includes a visit to an experimental field.
Coordinators
Coordinators
Dr. Marcel Bentancor, Dr. Esteban Casaretto, Dr. Gustavo Pereyra Alpuin
Teaching staff
Teaching staff
Dr. Jorge Prieto (INTA Mendoza, Universidad Juan Agustín Maza, Argentina) 🇦🇷
Dr. Thiago Teixeyra Santos (EMBRAPA digital, Brasil) 🇧🇷
Dr. Jonata Tyska Carvalho (Universidad Federal de Santa Catarina, Brasil) 🇧🇷
Dr. Leonardo Warzea Lima (Donald Danforth Plant Science Center, United States of America) 🇺🇲
Dr. Marcel Bentancor (Facultad de Ciencias, Universidad de la República, Uruguay) 🇺🇾
Dr. Omar Borsani (Facultad de Agronomía, Universidad de la República, Uruguay) 🇺🇾
Dr. Esteban Casaretto (Facultad de Agronomía, Universidad de la República, Uruguay) 🇺🇾
Dr. Gustavo Pereyra Alpuin (Facultad de Agronomía, Universidad de la República, Uruguay) 🇺🇾
Dr. Ignacio Ramírez Paulino (Facultad de Ingeniería, Universidad de la República, Uruguay) 🇺🇾
Number of students
Number of students
The course has 16 spots available for students enrolled through CABBIO: 3 for students from Brazil, 3 for students from Argentina, 8 for students from Uruguay, 1 for students from Paraguay, and 1 for students from Colombia.
The target audience for this course includes postgraduate students related to the following topics: Biotechnology, Agronomy, Engineering, Biology. Advanced students in these areas should justify their interest in participating. As this is an introductory course, no prior knowledge of machine learning is required. Familiarity with the topics to be covered in the course will be appreciated. In the case of postgraduate students, preference will be given to those associated with plant phenotyping projects.
Contact
Contact
If you want more information or have any questions, you can do so by contacting us at espacioprototipado@gmail.com.
Support:
Page updated
Report abuse