Machine learning models for assessing biodeterioration aspects of urban trees

• Grant number: 22/16562-7
• Support Opportunities: Scholarships in Brazil - Master
• Effective date (Start): March 01, 2023
• Effective date (End): April 30, 2024
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: João Paulo Papa
• Grantee: Giovani Candido
• Host Institution: Faculdade de Ciências (FC). Universidade Estadual Paulista (UNESP). Campus de Bauru. Bauru , SP, Brazil
• Associated research grant: 17/50343-2 - Institutional development plan in the area of digital transformation: advanced manufacturing and smart and sustainable cities (PDIp), AP.PDIP

Abstract

The present mastership project is part of the "Institutional development plan in the area of digital transformation (PDIP): advanced manufacturing and smart cities", promoted by the Institute for Technological Research of Sao Paulo (IPT) (Research Grant #17/50343-2), through the research line entitled "Machine learning models for assessing biodeterioration aspects of urban trees".This research line intends to monitor urban trees to determine their conservation conditions and potential risk of falling. To do so, it plans to apply statistical and machine learning models to find the external aspects of trees that possibly lead to the internal deterioration of the trunk.Thus, the project tries to improve the quality of life of urban populations, providing adequate management of urban forests. That is the reason why the research proposal is related to smart cities. Moreover, among the benefits of this management, it is possible to say:- the decrease in the risk of trees falling in urban areas and the number of accidents caused by this event;- the resulting improvement of other factors related to green areas, such as the decrease in the number of floods, the preservation of fauna, and the reduction of atmospheric pollutants;The contributions to the PDIP consist of the study of the external aspects of trees that are directly related to the biodeterioration of the main stem, having in mind the decrease of considerable fieldwork. It is worth mentioning the analysis of discriminant factors that contribute to the disparity of the probability of rupture of trees with similar characteristics. In this context, data collected by the field team will be analyzed to investigate the internal deterioration of threes and establish a correlation to the external aspects of the trees, such as trunk diameter e signs of stem rotting, as well as the development of machine learning models that can predict the internal deterioration rate of the trunk by inputting external aspects of the trees.Furthermore, this project considers the analysis of historical data collected with field work having in mind the investigation of the relation of internal and external measurements of trees with their respective probability of rupture. Therefore, clustering algorithms will be used to gather similar samples and analyze their respective probability of rupture of reference.