{"id":{"repo_id":"trento","oai_identifier":"oai:iris.unitn.it:11572/450295"},"canonical_url":"https://search.dev.ndltd.org/etd/trento/oai:iris.unitn.it:11572/450295","repository":{"repo_id":"trento","name":"Università degli Studi di Trento","base_url":"https://iris.unitn.it/oai/request"},"display":{"title":"EnogisAI: An Artificial Intelligence Framework for Predictive Agronomics","abstract":"Ensuring the quality and health standards necessary to feed the growing population is a critical challenge. Smart farming technologies combined with artificial intelligence (AI) represent an opportunity to improve crop quality and yield while efficiently and sustainably using the available planetary resources. This thesis describes EnogisAI: a comprehensive framework for developing and adopting agronomic AI solutions in the field. EnogisAI encompasses all the critical aspects of an AI solution development, starting from the data integration processes through model definition, training, and validation to the development of Decision Support Systems (DSSs) for farmers and agronomists. EnogisAI outputs are integrated within the Enogis agronomic management web platform, used to manage one-fourth of the total Italian viticulture, along with other crops. Relevant external environmental and weather data sources (e.g., Sentinel-2, ECMWF forecast, ERA5 reanalysis, TINITALY/01 digital terrain model) are integrated and linked with Enogis agronomic in-field collected data. Data is ingested into the data platform and is served through API interfaces to the production ecosystem. Machine and deep learning models are developed to address relevant agronomic challenges for Enogis users. Specifically, in this work, we describe (i) machine learning models for field zone identification for homogeneous management and yield prediction (R2>0.76), (ii) deep learning models for grapevine development monitoring using remote sensing (R2=0.96) and in-field images, (iii) a soil water balance assessment tools for crop stress detection and irrigation planning, and (iv) machine and deep learning models for reconstructing and forecasting pollen concentration time series (R2=0.82 and R2=.72, respectively). Finally, based on the performance and technological maturity levels, a selection of the resulting models is further developed into agronomic DSSs for crop development monitoring, disease protection, and soil water assessment. The DSSs are computed and updated daily for 121296 agricultural fields. This way, model results are accessible by end users through various tools, such as the Enogis WebGIS platform and field management tools, increasing the impact on sustainable agricultural practices. EnogisAI is the result of industrial research and innovation efforts. As such, all methodologies are designed and built based on the needs resulting from interacting with end users and stakeholders and in collaboration with technological patterns leaders in the agrometeorology ecosystem and international research experts. As an additional result, adopting an open innovation approach, a consistent part of the presented thesis is available either as conference proceedings or research articles.","abstract_html":"Ensuring the quality and health standards necessary to feed the growing population is a critical challenge. Smart farming technologies combined with artificial intelligence (AI) represent an opportunity to improve crop quality and yield while efficiently and sustainably using the available planetary resources. This thesis describes EnogisAI: a comprehensive framework for developing and adopting agronomic AI solutions in the field. EnogisAI encompasses all the critical aspects of an AI solution development, starting from the data integration processes through model definition, training, and validation to the development of Decision Support Systems (DSSs) for farmers and agronomists. EnogisAI outputs are integrated within the Enogis agronomic management web platform, used to manage one-fourth of the total Italian viticulture, along with other crops. Relevant external environmental and weather data sources (e.g., Sentinel-2, ECMWF forecast, ERA5 reanalysis, TINITALY/01 digital terrain model) are integrated and linked with Enogis agronomic in-field collected data. Data is ingested into the data platform and is served through API interfaces to the production ecosystem. Machine and deep learning models are developed to address relevant agronomic challenges for Enogis users. Specifically, in this work, we describe (i) machine learning models for field zone identification for homogeneous management and yield prediction (R2>0.76), (ii) deep learning models for grapevine development monitoring using remote sensing (R2=0.96) and in-field images, (iii) a soil water balance assessment tools for crop stress detection and irrigation planning, and (iv) machine and deep learning models for reconstructing and forecasting pollen concentration time series (R2=0.82 and R2=.72, respectively). Finally, based on the performance and technological maturity levels, a selection of the resulting models is further developed into agronomic DSSs for crop development monitoring, disease protection, and soil water assessment. The DSSs are computed and updated daily for 121296 agricultural fields. This way, model results are accessible by end users through various tools, such as the Enogis WebGIS platform and field management tools, increasing the impact on sustainable agricultural practices. EnogisAI is the result of industrial research and innovation efforts. As such, all methodologies are designed and built based on the needs resulting from interacting with end users and stakeholders and in collaboration with technological patterns leaders in the agrometeorology ecosystem and international research experts. As an additional result, adopting an open innovation approach, a consistent part of the presented thesis is available either as conference proceedings or research articles.","abstract_has_math":false,"creators":["Coviello, Luca"],"institution":"Università degli studi di Trento","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Brunato, Mauro"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025-04-08","date_published":"2025-04-08","updated_at":"2026-07-24T05:04:12Z","subjects":["agriculture, smart farming, artificial intelligence, decision support systems"],"languages":["eng"],"rights":["info:eu-repo/semantics/openAccess","license:Tutti i diritti riservati (All rights reserved)","license uri:iris.PRI01"],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["http://dx.doi.org/10.15168/11572_450295","10.15168/11572_450295"],"render_values":[{"text":"http://dx.doi.org/10.15168/11572_450295","href":"http://dx.doi.org/10.15168/11572_450295","code":true},{"text":"10.15168/11572_450295","href":"https://doi.org/10.15168/11572_450295","code":true}]}]},"links":{"outbound_url":"https://hdl.handle.net/11572/450295","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Coviello, Luca","Brunato, Mauro"]},{"key":"dc:creator","label":"Author","values":["Coviello, Luca"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2025-04-08"]},{"key":"dc:publisher","label":"Institution","values":["Università degli studi di Trento","place:TRENTO"]},{"key":"dc:relation","label":"Dc Relation","values":["firstpage:1","lastpage:146","numberofpages:146"]},{"key":"dc:type","label":"Dc Type","values":["info:eu-repo/semantics/doctoralThesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["agriculture, smart farming, artificial intelligence, decision support systems"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["info:eu-repo/semantics/openAccess","license:Tutti i diritti riservati (All rights reserved)","license uri:iris.PRI01"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://hdl.handle.net/11572/450295","http://dx.doi.org/10.15168/11572_450295","10.15168/11572_450295"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Ensuring the quality and health standards necessary to feed the growing population is a critical challenge. Smart farming technologies combined with artificial intelligence (AI) represent an opportunity to improve crop quality and yield while efficiently and sustainably using the available planetary resources. This thesis describes EnogisAI: a comprehensive framework for developing and adopting agronomic AI solutions in the field. EnogisAI encompasses all the critical aspects of an AI solution development, starting from the data integration processes through model definition, training, and validation to the development of Decision Support Systems (DSSs) for farmers and agronomists. EnogisAI outputs are integrated within the Enogis agronomic management web platform, used to manage one-fourth of the total Italian viticulture, along with other crops. Relevant external environmental and weather data sources (e.g., Sentinel-2, ECMWF forecast, ERA5 reanalysis, TINITALY/01 digital terrain model) are integrated and linked with Enogis agronomic in-field collected data. Data is ingested into the data platform and is served through API interfaces to the production ecosystem. Machine and deep learning models are developed to address relevant agronomic challenges for Enogis users. Specifically, in this work, we describe (i) machine learning models for field zone identification for homogeneous management and yield prediction (R2>0.76), (ii) deep learning models for grapevine development monitoring using remote sensing (R2=0.96) and in-field images, (iii) a soil water balance assessment tools for crop stress detection and irrigation planning, and (iv) machine and deep learning models for reconstructing and forecasting pollen concentration time series (R2=0.82 and R2=.72, respectively). Finally, based on the performance and technological maturity levels, a selection of the resulting models is further developed into agronomic DSSs for crop development monitoring, disease protection, and soil water assessment. The DSSs are computed and updated daily for 121296 agricultural fields. This way, model results are accessible by end users through various tools, such as the Enogis WebGIS platform and field management tools, increasing the impact on sustainable agricultural practices. EnogisAI is the result of industrial research and innovation efforts. As such, all methodologies are designed and built based on the needs resulting from interacting with end users and stakeholders and in collaboration with technological patterns leaders in the agrometeorology ecosystem and international research experts. As an additional result, adopting an open innovation approach, a consistent part of the presented thesis is available either as conference proceedings or research articles."]},{"key":"dc:title","label":"Title","values":["EnogisAI: An Artificial Intelligence Framework for Predictive Agronomics"]}]}],"canonical_facts":{"dc:contributor":["Coviello, Luca","Brunato, Mauro"],"dc:creator":["Coviello, Luca"],"dc:date":["2025-04-08"],"dc:description":["Ensuring the quality and health standards necessary to feed the growing population is a critical challenge. Smart farming technologies combined with artificial intelligence (AI) represent an opportunity to improve crop quality and yield while efficiently and sustainably using the available planetary resources. This thesis describes EnogisAI: a comprehensive framework for developing and adopting agronomic AI solutions in the field. EnogisAI encompasses all the critical aspects of an AI solution development, starting from the data integration processes through model definition, training, and validation to the development of Decision Support Systems (DSSs) for farmers and agronomists. EnogisAI outputs are integrated within the Enogis agronomic management web platform, used to manage one-fourth of the total Italian viticulture, along with other crops. Relevant external environmental and weather data sources (e.g., Sentinel-2, ECMWF forecast, ERA5 reanalysis, TINITALY/01 digital terrain model) are integrated and linked with Enogis agronomic in-field collected data. Data is ingested into the data platform and is served through API interfaces to the production ecosystem. Machine and deep learning models are developed to address relevant agronomic challenges for Enogis users. Specifically, in this work, we describe (i) machine learning models for field zone identification for homogeneous management and yield prediction (R2>0.76), (ii) deep learning models for grapevine development monitoring using remote sensing (R2=0.96) and in-field images, (iii) a soil water balance assessment tools for crop stress detection and irrigation planning, and (iv) machine and deep learning models for reconstructing and forecasting pollen concentration time series (R2=0.82 and R2=.72, respectively). Finally, based on the performance and technological maturity levels, a selection of the resulting models is further developed into agronomic DSSs for crop development monitoring, disease protection, and soil water assessment. The DSSs are computed and updated daily for 121296 agricultural fields. This way, model results are accessible by end users through various tools, such as the Enogis WebGIS platform and field management tools, increasing the impact on sustainable agricultural practices. EnogisAI is the result of industrial research and innovation efforts. As such, all methodologies are designed and built based on the needs resulting from interacting with end users and stakeholders and in collaboration with technological patterns leaders in the agrometeorology ecosystem and international research experts. As an additional result, adopting an open innovation approach, a consistent part of the presented thesis is available either as conference proceedings or research articles."],"dc:identifier":["https://hdl.handle.net/11572/450295","http://dx.doi.org/10.15168/11572_450295","10.15168/11572_450295"],"dc:language":["eng"],"dc:publisher":["Università degli studi di Trento","place:TRENTO"],"dc:relation":["firstpage:1","lastpage:146","numberofpages:146"],"dc:rights":["info:eu-repo/semantics/openAccess","license:Tutti i diritti riservati (All rights reserved)","license uri:iris.PRI01"],"dc:subject":["agriculture, smart farming, artificial intelligence, decision support systems"],"dc:title":["EnogisAI: An Artificial Intelligence Framework for Predictive Agronomics"],"dc:type":["info:eu-repo/semantics/doctoralThesis"]},"updated_at":"2026-07-24T05:04:12Z"}