Determinação de maturação em soja utilizando modelos multiplicativos e análise de imagens

• Autor(a) principal: Arantes, Pablo de Sousa
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFLA
• Texto Completo: http://repositorio.ufla.br/jspui/handle/1/56651
• topic: Glycine max (L.) Merrill; Maturação absoluta; Fenotipagem de alto rendimento; Soja - Florescimento; Absolute maturation; High-throughput phenotyping; Produção e Beneficiamento de Sementes

The adaptability of soybean cultivars could changes as the latitude, mainly due to the sensitivity of the crop to the photoperiod. Climate conditions can directly influence the growth and development of the soybean crop. Thus, soybean cultivars show variations in terms of full maturity in different sites, crop season, and sowing times. Thus, the purpose was to estimate the relative maturity of soybean cultivars using different strategies, as well as to measure the relative maturity prediction accuracy of the different methods. The experiments were carried out at the Centro de Desenvolvimento Científico e Tecnologico of the Federal University of Lavras - Fazenda Muquém, in Lavras - MG, located at latitude 21º12 S, longitude 44º58' W and altitude of 954 m, in the crop season, 2016/2017, 2017/18, 2018/19, 2019/20 and 2021/22. At Centro de Desenvolvimento e Transferência de Tecnologia – Fazenda Palmital, in Ijaci, located at latitude 21º09' S, longitude 44º54' W and altitude of 920 m, in the 2019/20 and 2021/22 crop season. The treatments were 54 commercial soybean cultivars in six environments, in a randomized complete block design, with different numbers of repetitions, with the plots consisting of four rows of five meters, thus making up an unbalanced data set. The evaluated trait was full maturity (days). Joint analysis (multi-environment), factor analyzes multiplicative mixed models (FAMM) and high-throughput phenotyping using UAVs were adopted. Statistical analyzes of the data were done using the R software. A regression model of the means for full maturity to estimate the relative maturity was obtained, adopting the stable cultivars. The FAMM method appears to be an efficient strategy for predicting full and relative maturity in soybean under tropical conditions. The use of image analysis by means of UAVs presents a lower correlation between the predicted RM and the original reported in the (National Cultivar Register), however, it is a promising strategy. The main challenge is the accuracy associated to the image uptake.