{"id":{"repo_id":"trento","oai_identifier":"oai:iris.unitn.it:11572/364650"},"canonical_url":"https://search.dev.ndltd.org/etd/trento/oai:iris.unitn.it:11572/364650","repository":{"repo_id":"trento","name":"Università degli Studi di Trento","base_url":"https://iris.unitn.it/oai/request"},"display":{"title":"Advanced Techniques For Prediction of Forest Above Ground Biomass Using Satellite Remote Sensing Data","abstract":"Large-scale mapping of forest AGB is a challenging task and crucial for forest management and planning. The use of SRS data has recently increased for AGB prediction due to their large footprint and low cost availability. There are various limitations and problems with SRS data that require innovative and effective solutions for large-scale AGB mapping. This thesis provides three main contributions in the context of using SRS for AGB prediction. The first contribution of the thesis involves the use of SMS data characterized by different spectral specifications, spatial resolutions and temporal availability. A systematic framework involving an adaptive regularization method was implemented to observe and quantify the response linked to various characteristics of the SMS data. The second contribution presents a dynamic generative neural network architecture for modelling AGB using multi-sensor satellite RS data. It proposes a method to derive AGB-oriented features and provides a seamless multi-sensor feature fusion method for AGB prediction. The third contribution presents a framework developed from the combination of a hyperparameter optimization procedure and a meta-learning algorithm to set up an end-to-end automated pipeline for modelling AGB. The contribution focuses on automatic development and extraction of features from MS data as well as automatic stacking of algorithms to compose an optimal ensemble model. The comprehensive analysis for each contribution was based on quantitative and qualitative results from the performed experiments. The first contribution pin-pointed the strengths and shortcomings of SMS data for AGB prediction in terms of effective spectral channels, effect of temporal information and role of spatial resolution for AGB prediction. The second contribution demonstrates the effectiveness of the proposed generative process in producing features that deliver more accurate AGB predictions as compared to the conventional approaches. Lastly, the third contribution generated automated features and stacked ensemble of models that outperformed individual models. The systematic series of experiments and flow of studies confirm that SRS data can be effectively used for accurately modelling AGB using the advanced methods demonstrated in the thesis.","abstract_html":"Large-scale mapping of forest AGB is a challenging task and crucial for forest management and planning. The use of SRS data has recently increased for AGB prediction due to their large footprint and low cost availability. There are various limitations and problems with SRS data that require innovative and effective solutions for large-scale AGB mapping. This thesis provides three main contributions in the context of using SRS for AGB prediction. The first contribution of the thesis involves the use of SMS data characterized by different spectral specifications, spatial resolutions and temporal availability. A systematic framework involving an adaptive regularization method was implemented to observe and quantify the response linked to various characteristics of the SMS data. The second contribution presents a dynamic generative neural network architecture for modelling AGB using multi-sensor satellite RS data. It proposes a method to derive AGB-oriented features and provides a seamless multi-sensor feature fusion method for AGB prediction. The third contribution presents a framework developed from the combination of a hyperparameter optimization procedure and a meta-learning algorithm to set up an end-to-end automated pipeline for modelling AGB. The contribution focuses on automatic development and extraction of features from MS data as well as automatic stacking of algorithms to compose an optimal ensemble model. The comprehensive analysis for each contribution was based on quantitative and qualitative results from the performed experiments. The first contribution pin-pointed the strengths and shortcomings of SMS data for AGB prediction in terms of effective spectral channels, effect of temporal information and role of spatial resolution for AGB prediction. The second contribution demonstrates the effectiveness of the proposed generative process in producing features that deliver more accurate AGB predictions as compared to the conventional approaches. Lastly, the third contribution generated automated features and stacked ensemble of models that outperformed individual models. The systematic series of experiments and flow of studies confirm that SRS data can be effectively used for accurately modelling AGB using the advanced methods demonstrated in the thesis.","abstract_has_math":false,"creators":["Naik, Parth Rajubhai"],"institution":"Università degli studi di Trento","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Bruzzone, Lorenzo"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2023,"date_issued":"2023-01-19","date_published":"2023-01-19","updated_at":"2026-07-24T05:04:17Z","subjects":[],"languages":["eng"],"rights":["info:eu-repo/semantics/openAccess","license:Creative commons","license uri:http://creativecommons.org/licenses/by/4.0/"],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["http://dx.doi.org/10.15168/11572_364650","10.15168/11572_364650"],"render_values":[{"text":"http://dx.doi.org/10.15168/11572_364650","href":"http://dx.doi.org/10.15168/11572_364650","code":true},{"text":"10.15168/11572_364650","href":"https://doi.org/10.15168/11572_364650","code":true}]}]},"links":{"outbound_url":"https://hdl.handle.net/11572/364650","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Naik, Parth Rajubhai","Bruzzone, Lorenzo"]},{"key":"dc:creator","label":"Author","values":["Naik, Parth Rajubhai"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2023-01-19"]},{"key":"dc:publisher","label":"Institution","values":["Università degli studi di Trento","place:TRENTO"]},{"key":"dc:relation","label":"Dc Relation","values":["firstpage:1","lastpage:149","numberofpages:149"]},{"key":"dc:type","label":"Dc Type","values":["info:eu-repo/semantics/doctoralThesis"]}]},{"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:Creative commons","license uri:http://creativecommons.org/licenses/by/4.0/"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://hdl.handle.net/11572/364650","http://dx.doi.org/10.15168/11572_364650","10.15168/11572_364650"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Large-scale mapping of forest AGB is a challenging task and crucial for forest management and planning. The use of SRS data has recently increased for AGB prediction due to their large footprint and low cost availability. There are various limitations and problems with SRS data that require innovative and effective solutions for large-scale AGB mapping. This thesis provides three main contributions in the context of using SRS for AGB prediction. The first contribution of the thesis involves the use of SMS data characterized by different spectral specifications, spatial resolutions and temporal availability. A systematic framework involving an adaptive regularization method was implemented to observe and quantify the response linked to various characteristics of the SMS data. The second contribution presents a dynamic generative neural network architecture for modelling AGB using multi-sensor satellite RS data. It proposes a method to derive AGB-oriented features and provides a seamless multi-sensor feature fusion method for AGB prediction. The third contribution presents a framework developed from the combination of a hyperparameter optimization procedure and a meta-learning algorithm to set up an end-to-end automated pipeline for modelling AGB. The contribution focuses on automatic development and extraction of features from MS data as well as automatic stacking of algorithms to compose an optimal ensemble model. The comprehensive analysis for each contribution was based on quantitative and qualitative results from the performed experiments. The first contribution pin-pointed the strengths and shortcomings of SMS data for AGB prediction in terms of effective spectral channels, effect of temporal information and role of spatial resolution for AGB prediction. The second contribution demonstrates the effectiveness of the proposed generative process in producing features that deliver more accurate AGB predictions as compared to the conventional approaches. Lastly, the third contribution generated automated features and stacked ensemble of models that outperformed individual models. The systematic series of experiments and flow of studies confirm that SRS data can be effectively used for accurately modelling AGB using the advanced methods demonstrated in the thesis."]},{"key":"dc:title","label":"Title","values":["Advanced Techniques For Prediction of Forest Above Ground Biomass Using Satellite Remote Sensing Data"]}]}],"canonical_facts":{"dc:contributor":["Naik, Parth Rajubhai","Bruzzone, Lorenzo"],"dc:creator":["Naik, Parth Rajubhai"],"dc:date":["2023-01-19"],"dc:description":["Large-scale mapping of forest AGB is a challenging task and crucial for forest management and planning. The use of SRS data has recently increased for AGB prediction due to their large footprint and low cost availability. There are various limitations and problems with SRS data that require innovative and effective solutions for large-scale AGB mapping. This thesis provides three main contributions in the context of using SRS for AGB prediction. The first contribution of the thesis involves the use of SMS data characterized by different spectral specifications, spatial resolutions and temporal availability. A systematic framework involving an adaptive regularization method was implemented to observe and quantify the response linked to various characteristics of the SMS data. The second contribution presents a dynamic generative neural network architecture for modelling AGB using multi-sensor satellite RS data. It proposes a method to derive AGB-oriented features and provides a seamless multi-sensor feature fusion method for AGB prediction. The third contribution presents a framework developed from the combination of a hyperparameter optimization procedure and a meta-learning algorithm to set up an end-to-end automated pipeline for modelling AGB. The contribution focuses on automatic development and extraction of features from MS data as well as automatic stacking of algorithms to compose an optimal ensemble model. The comprehensive analysis for each contribution was based on quantitative and qualitative results from the performed experiments. The first contribution pin-pointed the strengths and shortcomings of SMS data for AGB prediction in terms of effective spectral channels, effect of temporal information and role of spatial resolution for AGB prediction. The second contribution demonstrates the effectiveness of the proposed generative process in producing features that deliver more accurate AGB predictions as compared to the conventional approaches. Lastly, the third contribution generated automated features and stacked ensemble of models that outperformed individual models. The systematic series of experiments and flow of studies confirm that SRS data can be effectively used for accurately modelling AGB using the advanced methods demonstrated in the thesis."],"dc:identifier":["https://hdl.handle.net/11572/364650","http://dx.doi.org/10.15168/11572_364650","10.15168/11572_364650"],"dc:language":["eng"],"dc:publisher":["Università degli studi di Trento","place:TRENTO"],"dc:relation":["firstpage:1","lastpage:149","numberofpages:149"],"dc:rights":["info:eu-repo/semantics/openAccess","license:Creative commons","license uri:http://creativecommons.org/licenses/by/4.0/"],"dc:title":["Advanced Techniques For Prediction of Forest Above Ground Biomass Using Satellite Remote Sensing Data"],"dc:type":["info:eu-repo/semantics/doctoralThesis"]},"updated_at":"2026-07-24T05:04:17Z"}