Computational platform for botanic identification through dendrological characteristics

• Grant number: 18/01036-2
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: July 01, 2019
• End date: June 30, 2020
• Field of knowledge: Biological Sciences - Botany - Plant Taxonomy
• Principal Investigator: Ana Catalan
• Grantee: Ana Catalan
• CNAE: Atividades profissionais, científicas e técnicas não especificadas anteriormente
• Associated scholarship(s): 19/15385-1 - Computational platform for botanic identification through dendrological characteristics, BP.TT; 19/14250-5 - Computational platform for botanic identification through dendrological characteristics, BP.TT; 19/14132-2 - Computational platform for botanic identification through dendrological characteristics, BP.PIPE

Abstract

The aim of the project is to develop a botanical identification system, based on computer visualization and machine learning, focusing on the environmental licensing market, academic research and biotechnology area. The development of a methodology that allows standardizing data collection in the field and at the same time quickly and accurately identify the raised plant species, will provide a useful tool for the professional, reducing the botanical identification time. Once the data collection is done, the system stores the information and estimates the correlations between the different vegetative characters of the specimens, grouping characteristics of the main botanical families, ordering the species in hierarchical categories according to both natural affinities or degrees of kinship and with a dichotomous classification system. It is intended to create a computational platform that, through the inclusion of data and photos on the reproductive, vegetative and organoleptic characteristics (smell, texture, etc.) of the trees, recognize of the species as identical or similar to another already known and in (family, genus and species), considering the botanical families of greatest interest in the study region, facilitating the identification of the trees in the field, leading to more precise results without the need to make exsiccation. The overcoming of technical challenges related to the collection of quality information in loco, as well as the pattern recognition, will have a direct impact on the quality and accuracy of the results found in the environmental reports, as well as on the reduction of the time and specialized human resources involved in the process. This will give a good differential to the aimed solution, making it more attractive to the market. The methodology used includes the development of a prototype, processing tests to validate the main ideas, to increase the research on solutions offered in the market and elaborating the business plan. The expected result at the end of Phase I is the demonstration of the technical and economic feasibility in relation to the application of the solution proposed in the botanical identification process. (AU)