Machine learning methods in remote sensing for forest disturbance ecology, Koli NP Finland, Part 1 скачать в хорошем качестве

Machine learning methods in remote sensing for forest disturbance ecology, Koli NP Finland, Part 1

Instagram:   / resdinet   Facebook:   / resdinet   *********************************************************** The second RESDiNET Workshop in the use of machine learning technologies in remote sensing for forest disturbance ecology, Feb 12-16, 2024, Koli NP, Finland. Dr. Raquel Alves de Oliveira and Dr. Roope Näsi from the Finnish Geospatial Research Institute provide hands-on training in using the Random Forest algorithm to classify trees based on their status (healthy, bark beetle-infested, dead) using spectral features and vegetation indices derived or calculated from tree canopy spectral reflectance. The analysis is conducted in a Python programming environment. *********************************************************** RESDiNET is an international research project funded by the European Research Executive Agency, Horizon Europe. The project enhances networking activities between research institution in Widening country (IFE SAS) and top-class counterparts at the EU level (Finnish Geospatial Research Institute, The University of Eastern Finland and Swedish University of Agricultural Sciences). The project builds on networking for excellence through knowledge transfer and exchange of best practices between involved institutions. The project establishes initial network and development of a new joint research project in novel RST applications in FDE. Rigorous analyses of severe insect-induced disturbances using novel RST is carried out in test areas representing different forest and climate types: mountain forests in Slovakia and boreal forests in Finland and Sweden. The project integrates in situ UAV and drone acquired remotely sensed data, existing multitemporal geospatial information and field data, particularly data on bark beetle population density, visible infestation symptoms linked to outbreak phases, and tree physiology parameters measured using electronic dendrometers or sapflow meters. The combined dataset are used to develop new tools for landscale-level early bark beetle attack identification and for designing bark beetle infestation risk assessment model. We draw on the latest advances in drone technologies and image analytical tools, including deep Convolutional Neural Networks based machine learning techniques and Artificial Intelligence algorithms. machinelearning #remotesensing #python #randomforests #forestdisturbance #forestdisturbanceecology #multispectralimagery #barkbeetle #barkbeetleoutbreak #remotelandmonitoring #earlydetection #barkbeetleinfestation #norwayspruce #barkbeetledisturbance #climatechange #drought #ipstypographus #piceaabies #treemortality #researchproject #horizoneurope #horizonwidera #europeanresearch #forestecology #forestmanagement #finland #uef #coniferousforests