Fuzzy Power, Converter for Plastic Bottle Sorting System Based on Weight

Abstract

Presently, intelligent robot technology has pervaded various realms of human activities and productivity. The future trajectory of development necessitates a shift toward employing intelligent robots for automatic sorting, replacing human involvement. Within the domain of sorting robots, those dedicated to bottle sorting hold significant research importance. This paper focuses on a complex machine vision (MV) sorting application involving plastic bottle identification. The inherent challenge lies in the difficulty of distinguishing between bottle types due to the absence of distinct features and the variability within each type. Although existing literature documents bottle recognition algorithms, they are primarily evaluated offline using bottle weight. This paper evaluates an MV-based system designed for online recognition and enumeration of plastic bottles as they traverse a conveyor belt. The study assesses the accuracy and efficacy of three weight-based techniques. Ultimately, none of these techniques yielded satisfactory outcomes, failing to achieve the targeted 95% accuracy for 1000 bottles.