Analysis of Biodegradable and Non-Biodegradable Materials Using Selected Deep Learning Algorithms

It is possible to divide the materials used in the world into recyclable and nonrecyclable. Biodegradable materials contain elements naturally degraded by microorganisms such as foods, plants, fruits, etc. Waste from this material can be processed into compost. non-biodegradable materials include materials that do not naturally decompose, such as plastics, metals, inorganic elements, etc. Waste from this material can only be reused by converting it into new materials. In this study, the classification of biodegradable and non-biodegradable materials was done using deep learning methods. Convolutional Neural Network (CNN) performs steps such as preprocessing and feature extraction in classification. 5430 images were used for the dataset. 70% of this dataset was used as training data, 15% as validation data, and 15% as test data. Of the Deep Learning methods, the pre-trained neural networks AlexNet, ShuffleNet, SqueezeNet, and GoogleNet were used. For each algorithm, the performances were evaluated by classifying them as biodegradable and non-biodegradable. With this study, we can identify, track, sort, and process waste materials by classifying materials