Soft-Sensor based on $k$-NN classifiers in Agricultural Pest Control

摘要

The quality of agricultural sprays plays an important role in the application of chemical products dedicated to plant protection and food production. In such context, the effects of nozzle type, standard and size, as well as the operating pressure should be well observed. Measured droplet sizes and velocities are also important, and both can be affected by nozzle type, size, and operating pressure. Therefore, control aspects have become necessary to improve quality and to reduce undesirable externalities in the agricultural environment. This paper, presents a soft-sensor based on the $k$-NN classifier and estimator, which requires an attribute vector with the spraying quality descriptors, and delivers an estimation for the best values of the descriptors to obtain an adequate quality in relation to the spraying in agricultural pest control. The results of this research reveal the usefulness of such soft-sensor architecture based on estimation theory and added value in the application of agricultural pesticides.