O reconhecimento de espécies arbóreas em campo por meio da casca com o uso da espectroscopia no visível e infravermelho próximo na Amazônia Central
The recognition of the identity of tree plants is a subjective process that presents many difficulties, especially in a highly diverse environment such as the Amazon. The problems caused by misidentification affect from botanical inventories to the monitoring and commercialization of wood as the fin...
| Autor principal: | Hadlich, Hilana Louise |
|---|---|
| Grau: | Dissertação |
| Idioma: | por |
| Publicado em: |
Instituto Nacional de Pesquisas da Amazônia - INPA
2020
|
| Assuntos: | |
| Acesso em linha: |
https://repositorio.inpa.gov.br/handle/1/5168 http://lattes.cnpq.br/2228944258807319 |
| Resumo: |
|---|
|
The recognition of the identity of tree plants is a subjective process that presents many difficulties, especially in a highly diverse environment such as the Amazon. The problems caused by misidentification affect from botanical inventories to the monitoring and commercialization of wood as the final product. Previous studies have shown that near infrared spectroscopy is effective for discriminating tree species using both leaf and wood. However, these studies were performed in the laboratory and with dry material, necessitating the determination of methods under field conditions. In order to improve the forest inventory system in Amazonia and to assist in the recognition of species for management, a combination of portable equipment and a spectral library is necessary. Therefore, this study had objective to test the visible and near infrared spectroscopy technique in field on trunk of trees, to recognize the species through spectra of the bark tissues (rhytidome and phloem). The spectral collection was made in 11 Amazonian species with the ASD field spectrometer covering the region of visible to near infrared. This technique proved to be very efficient because it got a hit level of 98% of species recognition using the inner bark (phloem) and 94% accuracy for the outer bark (rhytidome) with the average spectra. As is a field technique, it was also tested whether the moisture in the bark influences the species recognition. Using data collected from four species in two distinct periods and discriminant models, the function that used the data from both periods with the inner bark was the one that best predicted the species (97%). That is, we can collect spectra from the trees at any time of the year without significantly influencing ecognition. This technology has been successful in the field, and could be used to give more reliability in the process of identification of inventories, assisting to conserve biodiversity. |