GARBAGE CLASSIFICATION USING DEEP LEARNING

Garbage disposal is a critical responsibility in maintaining a healthy and green environment. The recycling business is thriving as inhabitants of India become more aware of the necessity of a clean environment in terms of reducing natural resource use and rubbish disposal. The amount of trash generated in daily life has an impact on land, water, and air, posing a major threat to aquatic animals and their habitats, as well as humans, if not properly managed. Conventional garbage disposal systems are on the rise, requiring accurate and efficient segmentation and recognition processes. This demand corresponds to an increase in the computing capacity of modern computer systems as well as more efficient picture recognition methods. To address this issue, the garbage categorization process is automated by using an image classifier with a convolutional neural network (CNN), which reduces waste segregation time and costs. The goal of automating the process is to reduce human intervention and increase productivity in the waste segregation process. In this paper, three distinct models are tested for greater accuracy: Simple CNN, ResNet50, and VGG16 are trained on various image datasets and used to extract features from images, which are then fed into a classifier for dump/trash categorization. The learned models are then loaded into the mobile application.