Análise de tecnologia verde e inteligência artificial aplicadas à predição de atributos do solo na Amazônia brasileira

Abstract

The present study aims to explore these proximal sensors (pXRF and Vis-NIR) in predicting attributes of soils in the Amazon, specifically soils in the state of Pará, to contribute to the advancement of knowledge about the use of these sensors in tropical soils and adopt its use in laboratories. A state with a large territorial extension, with strong agricultural aptitude and growing demand from farmers for basic information on soil texture and fertility are motivations to overcome the obstacles of geographical distance from analysis laboratories. As this research begins with Amazonian soils, progress can be made in creating accurate models applicable to this region, making sensors an efficient tool for understanding regional soils, whether for agricultural production or for the preservation of these soils. in protected areas. Our hypothesis is that data from pXRF and Vis-NIR sensors combined are capable of predicting the texture and fertility of Amazonian soils with more accurate results than data from isolated sensors.