Use este identificador para citar ou linkar para este item: http://repositorio.ufla.br/jspui/handle/1/46104
Título: Monitoramento de características morfométricas e vigor de cultivares de café por ciência de Sensoriamento Remoto Multisensor
Autores: Alves, Marcelo de Carvalho
Alves, Marcelo de Carvalho
Resende, Mário Lúcio V.
Carvalho, Gladyston Rodrigues
Palavras-chave: Machine learning
Sensoriamento remoto
Cafeicultura
Índice vegetativo de diferença normalizada
Vegetative index of normalized difference
Remote sensing
Data do documento: 9-Fev-2021
Editor: Universidade Federal de Lavras
Citação: CAMPOS, G. A. de O. Monitoramento de características morfométricas e vigor de cultivares de café por ciência de Sensoriamento Remoto Multisensor. 2020. 57 p. Dissertação (Mestrado em Egenharia Agrícola) – Universidade Federal de Lavras, Lavras, 2021.
Resumo: With the use of remote sensing it was possible to obtain information at low flight altitudes and with very high spatial resolution, in coffee growing. The objective was to develop and evaluate techniques to monitor the morphometric characteristics and the vegetative vigor of the coffee tree with data obtained by direct and indirect measurements in situ, aerial monitoring by orthorectified RGB images at 2 flight altitudes and multispectral imaging of the Earth's orbit using bands 4 and 8 for red and infrared, respectively. The study was conducted in the experimental area of INCT do Café, at the Federal University of Lavras. In situ data collection was performed with the equipment, graduated topographic ruler, NDVI Greenseeker ™ sensor, and the leaf area index sensor, LAI2000 compared to aerial monitoring with UAS DJI Phantom 4 and, orbital, with the Sentinel 2A. Points were collected to obtain the plant height, crown diameter, vegetative index of normalized difference (NDVI) and leaf area index (LAI) of 20 arabica coffee cultivars from June to August 2019. Two aerial surveys by means of UAS at the height of 30 and 70 meters and for the same date that the imaging was carried out with Sentinel 2A, around 11:00 am The surface and terrain elevation model was calculated and, by the difference of the images, plant height was obtained. The LAI was calculated for data collected in situ and plant height and crown diameter extracted from UAS images. Regarding NDVI, Machine Learning algorithms by Random Forest and Multiple Linear Regression were used to calibrate the NDVI of UAS based on the NDVI of the plants measured in situ with the Greenseeker ™ and another model comparable to the NDVI obtained by imaging with Sentinel 2A. It was possible to determine an NDVI vigor value of UAS RGB to monitor the canopy of 20 coffee cultivars with an R² value equal to 0.72. With regard to plant crown morphometry, the height of the plants measured by UAS image processing was close to that determined in situ, with an average difference of less than 0.15m.
URI: http://repositorio.ufla.br/jspui/handle/1/46104
Aparece nas coleções:Engenharia Agrícola - Mestrado (Dissertações)



Os itens no repositório estão protegidos por copyright, com todos os direitos reservados, salvo quando é indicado o contrário.