Automated Garbage Waste Management using Deep Learning for sustainability

As smart cities evolve, waste management constitutes one of the major primary elements of it. Recycling can be made more affordable in the vital phase of garbage sorting process. However, categorizing garbage manually involves an enormous amount of effort and time. A machine vision technique is presented in this investigation to autonomously recognize and divide garbage into six categories—plastic, paper goods, glass, metallic materials, cardboard, and trash. There are over 2509 pictures in the deep learning structure dataset. Garbage datasets are also additionally collected and tested in real time employing mobile devices. The model is constructed using CNN-2D and Yolo Tiny V3, deep convolution neural network-based methods. Furthermore, dropout and data augmentation have been employed to enhance accuracy. The results indicate that the proposed Yolo Tiny V3 approach operates effectively in garbage classification, obtaining greatest likelihood distributions for every garbage class that was categorized.