Intelligent Based Waste Management Awareness Developed by Transfer Learning

Urban India produces 62 million tons of garbage (MSW) per year, with projections of 165 million tons by 2030. Annually, 43 million tons of urban solid wastes are collected, with 31 million being discarded in landfills and just 11.9 million being handled. In India, many people litter the highways as well. They probably just filled the road with banana leaves or dried leaf bowls a few years ago; those types of litter were not particularly dangerous because they were environmentally friendly and could even be eaten by abandoned animals. However, in today's India, plastic is the most commonly littered object and cultural change is difficult to achieve in any society. If it continues like this, India will become an unsuitable country for living. To overcome this crucial problem, our research will be the best solution. We analysis on nine different classes so that the recycling authority can recycle these easily and reduced the wastes. Our proposed model has been designed with three different tasks. In the beginning step, we have to capture any particular wastes. In the second step, we have to classify these captured images with the help of web applications that have been developed by the Transfer Learning model (VGG16) and according to the classification name, we have to put them into the same type bucket so that every bucket contains only one unique type of wastes. Finally, we have to call the authority who is involved with recycling so that they can easily recycle them and adequately reduce the waste. Our model accuracy is almost 96%. Even though these works are initialization stage done by the web application but in the future, we will develop a mobile application to make it easier and user friendly.