SSD based waste separation in smart garbage using augmented clustering NMS

Object detection plays a pivotal role in autonomous systems helps to build the machines to be intelligent as humans that leads to build an Artificial Intelligence application used for society,Industries,face-recognition and so-on. Nowadays, it is difficult to manage the waste generated by humans and industries which is increasing rapidly day by day needs to address the problem to make automation for separating the Bio-degradable and Non-Bio degradable waste. Although humans are tried to manage impact of waste management in society to maintain the eco-system by implementing a separate trash for Bio and Non-bio waste. Sometimes it is difficult to follow for the separation of waste manually by humans. There is no existing sensor to identify the types of wastes. In this proposed system, it evolves the implementation of the bio-degradable and non-bio degradable object detection method to help to detect these objects automatically with the augmented clustering NMS using Single-shot detector methods. The enhanced augmented clustering algorithm effectively detects the multiple objects in the video along with the respective bio or non-bio classification custom object detection model. With the build thousand images for each class custom dataset model to train the objects using deep learning neural network. A custom object detection data model is built with the help of NVIDIA GPU RTX 4 GB using tensorflow model. Here the results are interpreted with the mean average precision value of 0.965 with ACNMS Single shot object detector which is effectively detected with the new enhanced technique.