Establecimiento de modelos parametrizados para estimación de posible precencia de contaminación y enfermedad del suelo mediante drones en la zona costera

Abstract

The implementation of remote sensors to estimate the parameters of the quality of the soil has demonstrated to be an efficient tool to protect the soil and the prediction of fertile soil. This investigation´s analysis is done through the use of environmental indicators, the behavior of parameters such as N, P, MO and humidity, in order to evaluate changes and alterations of soil in the Coastal regions of Ecuador. Various linear regression models have been generated using spectral indices, among them three indices of vegetation the first of normalized difference (NDVI) and the second of normalized green difference (GNDVI), and the last one known as index normalized difference red edge (NDRE); in order to estimate the water stress of the soil, the standardized water index (NDWI) and the soil moisture index (SMI), where the one with greatest correlation with N is the GNDVI R2 = 0.7958, residual standard error (RSE) 0.03664, according to the NDRE (R2=0.7578), which is associated to the MO factor (R2 = 0.5425, RSE 0.5975), the NDVI attained a stronger relationship with the phosphorus (R2 = 0.6782, RSE 0.9983) in comparison to other parameters. Finally, the SMI had a better association with humidity (R2 = 0.7433, RSE 2.457) continued the NDWI is compared in order to determine the cultivation´s water content. In conclusion, the variables implemented are comparable to the nutritional requirements of the soil where an excess or deficit of one of these can destabilize the structure of the soil.