Raspberry Pi-based design of intelligent household classified garbage bin

Abstract

Until now, waste pollution is a global problem. The first and foremost part of the solution is waste classification, as it can effectively reduce the output of waste, so it is particularly important to classify household garbage in the household. However, classification of household garbage is a difficult task for people, and there are not enough designs and researches on automatic garbage classification bins at home. In this study, we designed and built a light and convenient intelligent garbage bin using devices such as motors, turn plates and a camera for classifying household garbage. Compared with the traditional garbage classification methods, two improvements in the algorithm are made, including the combination of the EfficientNetB2 model with the parallel mixed-attention mechanism and the design of the background noise removal algorithm. Finally, the algorithm is disposed to the Raspberry Pi to compose the complete classification system. The accuracy of this classification system in the test set was 93.38% and the results show that this bin can effectively distinguish recyclable waste, kitchen waste, hazardous waste and other waste, and is suitable for daily life garbage identification at home.