WASTE SEGREGATION MODEL USING ARTIFICIAL NEURAL NETWORKS

The three fundamental stages of the Waste Segregation System are: • Acquisition of the images of the waste material • Classification by the Convolutional Neural Network • The physical segregation of the waste is conducted by the means of a control mechanism. A conveyer belt system is used to realize the working of the Waste Segregation System. This belt is programmed to carry the waste to the camera module, whenever an object is detected at the input. This is achieved by using a laser and a photodetector as object detectors. Similarly, another object detector is used at the camera module, to detect the presence of an object, to capture an image of it. The captured image is then fed to the convolutional neural network, present in the Decision-Making Module, which outputs a label according to the category of waste that the the object is segregated into. The label thus generated is used to control the blocking arm present above the respective bin that the waste must fall into. The blocking arm is programmed to open out to a particular angle, effectively blocking the path of the object on the conveyer belt, and then close back to its initial position, hence performing a sweeping action. This combined motion of the belt and the arm will ensure that the waste falls into the respective bin.