AUTOMATIC WASTE SEGREGATION SYSTEM USINGCONVOLUTIONAL NEURAL NETWORKS

Automatic Waste Segregation is the process of classifying and separating waste objects into several categories for further treatment, it acts as a middle process between cumulative collection and waste treatment. Improper waste treatment not only causes pollution but also increases the overall cost of the waste treatment process, hence a proper waste segregation mechanism needs to be implemented before treatment. The paper is a successful attempt of creating a waste segregation mechanism with minimal cost and with minimal human intervention at household level with the help of deep learning specifically Convolutional Neural Network (CNN), image recognition and Internet of Things (IoT). The model classifies a waste as either biodegradable or non-biodegradable and segregates waste into separate compartments using a flap mechanism, the proposed model also sends feedback to the user in real time as when the bin requires maintenance. The neural network is trained using a large dataset consisting of approx. 13000 images combined with suitable hidden layers, 70-30 % images were used for training and testing respectfully. The proposed model gives accuracy of 92%. The classification and segregation happens in real time which helps to reduce overall human efforts and reduces cost of overall process.