Identificação de massas desbalanceadas em lavadoras de roupas com técnicas de redes neurais e visão computacional

Abstract

The spinning phase is a primary function of washing machines which has the objective of extracting residual water from clothes by spinning a vertical axis basket suspended in four points. Either by the washing process or by inappropriate consumer usage, when the clothes are not uniformly distributed there is an unbalancing phenomenon that can result from small hit noises to product destruction. The research has the objective of evaluating the application of neural network techniques to identify unbalanced mass in washer machines using collected data from a computer vision system. By observing variables such as angular speed and translational movement from different parts, artificial intelligence techniques will be applied in order to identify and classify patterns created by different intensities of unbalancing loads. Initially, the systems of a washing machine and the unbalancing phenomenon will be described. The bibliographic review of digital image processing supports the computational vision system and algorithm. The practical experimentation will collect a data set from the entire inference space of possible unbalancing configurations, in such a way that enough data is provided to train, validate and test the neural network tools. It is expected the proposed system be capable of identifying and classifying different intensities of an unbalanced load.