Análise da qualidade da aplicação georeferenciada de chuva artificial em área de plantio com visão computacional e rede neural.

Abstract

Agricultural monitoring on the fields is receiving attention and it has been emphasize the optimization of impacts that occur due the inadequate use of agricultural products. This paper presents a methodology for quality analysis of pesticide application in agricultural field based on both machine vision and artificial neural network. The technique uses digital image and Hough Transform processing for artificial raindrop counting and distribution, as well as a model for decision making with a neural network, which receives descriptors of the raindrop distribution for specific-site analysis.