Use of Unmanned aerial vehicle images as a tool to evaluate stand uniformity in clonal Eucalyptus plantations Ana Rosária Sclifó Zucon 1 Brandon Hawkes 2 Cristiane Camargo Zani de Lemos 3 Guilherme Zaghi Borges Batistuzzo 3 Rodrigo Eiji Hakamada 3 Guilherme Rodrigues de Pontes 3 Thiago Ubiratan de Freitas 3 José Henrique Bazani 1 Clayton Alcarde Alvares 4 José Carlos Arthur Junior 4 José Leonardo de Moraes Gonçalves 1 1 Universidade de São Paulo - USP/ESALQ Caixa Postal 9 - 13418-900 - Piracicaba - SP, Brasil zuconana@gmail.com bazani.jh@gmail.com jlmgonca@usp.br 2 North Carolina State University Raleigh, North Carolina, USA bahawkes@ncsu.edu 3 International Paper Rod. SP 340, Km 171 Mogi Guaçu – SP - 13840-970 cristiane.lemos@ipaper.com guilherme.batistuzzo@ipaper.com rodrigo.hakamada@ipaper.com guilherme.pontes@ipaper.com thiago.freitas@ipaper.com 4 Instituto de Pesquisas e Estudos Florestais – IPEF Avenida Pádua Dias, 11 - Caixa Postal 530 - CEP: 13400-970 - Piracicaba/SP arthur@ipef.br clayton@ipef.br Abstract. The monitoring of silvicultural quality in eucalyptus clonal plantations is essential to check the quality of forest operations. The uniformity of a stand is considered a fundamental variable to the monitoring. It shows potential improvements in silvicultural practices, enabling increases in stand uniformity and yield. Today, Unmanned Aerial Vehicle (UAV) images are underutilized and have been shown to be a powerful tool in forest monitoring. The main goal of this work was to develop a model to identify the tree’ crowns and create a uniformity evaluation method in clonal eucalyptus plantations through the use of UAV. 50 plots of 50m x 20m were allocated within the study area using the ArcGIS 10.1 software. The maximum likelihood classification tool was used, dividing the image in two classes, tree and non-tree. It was then transformed into polygons and the area of the crowns were calculated. The Pvar50 was used as a uniformity index for the crown areas. The result was a low uniformity index (31, 5%) and high fail index (10, 2%), which can be explained due to possible failures in forest activities or drought that occurred after the planting period. The evaluation of the UAV images proved to be a powerful tool to check the silvicultural quality of the stands. In order to get a good maximum likelihood classification, a high image quality it required. Key-words: UAV images, uniformity, model builder, silvicultural quality Anais XVII Simpósio Brasileiro de Sensoriamento Remoto - SBSR, João Pessoa-PB, Brasil, 25 a 29 de abril de 2015, INPE 4636