Systems and methods for optical material characterization of waste materials using machine learning

Abstract

Systems and methods for optical material characterization of waste materials using machine learning are provided. In one embodiment, a system comprises: an imaging device configured to generate image frames an area and target objects within the area; an object characterization processor coupled to the imaging device and comprising Neural Processing Units and a Neural Network Parameter Set. The Neural Network Parameter Set stores learned parameters utilized by the one or more Neural Processing Units for characterizing the one or more target objects. The Neural Processing Units are configured by the Neural Network Parameter Set to detect a presence of a plurality of different materials within the image frames based on a plurality of different features. For a first image frame of the plurality of image frames, the Neural Processing Units outputs material characterization data that identifies which of the plurality of different materials are detected in the first image frame.