Avaliação das condições hídricas em lavoura cafeeira por meio de geoestatística e aeronave remotamente pilotada

• Autor(a) principal: Santos, Sthéfany Airane dos
• Data de Publicação: 2021
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFLA
• Texto Completo: http://repositorio.ufla.br/jspui/handle/1/46533
• topic: Agricultura de precisão; Aeronaves remotamente pilotadas; Sensoriamento remoto; Irrigação agrícola; Café - Estresse hídrico; Precision agriculture; Remotely piloted aircraft; Remote sensing; Agricultural irrigation; Coffee - Water stress; Geoestatística; Geostatistics; Engenharia agrícola

The Remotely Piloted Aircraft (RPA’s) has become an important technology for Precision Agriculture (PA), in the last decade they allowed the acquisition of images by remote sensors, with high spatial and temporal resolution, in addition to providing different information. Therefore, the objective of this work was to evaluate the water conditions of a coffee plantation through geostatistics and the use of high-resolution images for the calculation of vegetation indices. This study was conducted in an area of 1.2 ha, under the cultivation of coffee trees of the species Coffea arabica L., cultivar Topázio MG 1190. The study area and the 30 sampling points were georeferenced using a GNSS RTK. Data collection was carried out in two seasons, dry period (August 2020) and rainy period (January 2021). High resolution images were obtained using an RPA equipped with a multispectral sensor, for the Red, Nir, Green and Red Edge bands. The flight plan was elaborated in the eMotion software and the images obtained were processed by the Pix4D software, which created a point cloud, a digital surface model and an orthomosaic of the images. 30 undisturbed soil samples were collected, at a depth of 0-10 cm and 30 samples at a depth of 10-20 cm, which later went through the drying process in an oven at 105ºC for 24 hours to establish the soil density, gravimetric moisture and volumetric humidity. Leaves were collected at 4:30 am in georeferenced plants, where water potential was determined by using a Scholander Pump. The spatialization and interpolation of data on soil moisture and leaf water potential was carried out by geostatistical analysis, with adjustment of semivariograms and creation of maps by ordinary kriging. Through the images obtained by the ARP, vegetation indices were calculated. From the correlation analysis and linear regression, it was verified the relation of the attributes obtained in the field and the vegetation indexes. The degree of spatial dependence obtained by the geostatistics data showed a strong spatial dependence for all evaluated attributes and for both years of collection. The vegetation indices showed a significant difference when comparing the dry and rainy periods. For the analysis of correlation between field data and vegetation indices, the highest value was between the GREEN spectral band and the volumetric humidity collected at a depth of 0-10 cm for the year 2020 (51.57%). The water potential of the leaves of 2021, correlated significantly with a spectral band and six vegetation indexes, whereas the linear regression that obtained the best fit was for the water potential attribute of 2021 with the NDRE index. These results show the efficiency of geostatistical tools and RPA, for the evaluation of water conditions, and that through even more in-depth studies, they can become great allies to coffee growing.