An Autonomous Garbage Removal Approach Using Resnet as Backbone and MaskRcnn for Segmentation

A safe and hygienic urban landscape has become more important than ever due to population growth and industrialization. In order to recycle garbage, sorting it is one of the most important steps in municipal waste management. Having to sort by hand is time-consuming, tedious, and costly. In order to improve garbage sorting, robots could be used for automated waste management implementations. Preliminary material separation, followed by detailed segregation, is an essential component of the whole process. Recent years have seen artificial intelligence and deep learning being used in a number of different fields. To automatically detect and locate garbage in real-world images and videos, this research proposes a Garbage Detection System based on object detection models. Reviewing previous research, the paper proposes a new method for detecting garbage. In this experimentation, we are proposing the Region-based Convolutional Neural Networks (Mask R-CNN) model for Object detection and segmentation which is built on Feature pyramid network(FPN) and ResNet50 as backbone architectures. Also, as part of this experimentation, we are planning to collect a real-time garbage dataset for training purpose.