Overview / Usage

The research is focussed on coming up with an efficient garbage sorting and monitoring system. The system is set to perform several tasks as stipulated below.
i. Sort the garbage in the dustbin into different categories; plastics, metals, paper bags etc.
ii. Monitor the garbage level
iii. Alert the necessary garbage collecting authorities whenever the dustbin is full
iv. Keep a well-documented record of the garbage collection schedule

Waste is thrown into the system and the system sigularizes the garbage into single entities in the Sorting system. The Separation system determines which material it is made from. The prototype distinguishes between three
different kinds of inorganic waste; aluminum cans, plastic bottles and glass. The
process to identify the waste is split into three categories; image processing, characteristic
extraction and machine learning. The output of each step is the input to the next step. First, the system starts when the button in the graphic user interface (GUI) represented by the touch screen is pressed. It triggers the
system and the process in the multimedia processor (MMP) begins. Second, the webcam
that is connected to the USB port takes an RGB image of the waste shape and sends
it to the MMP. The latter processes the image in order to get the features of the waste.
The system classies the target based on the rst two Hus Invariant Moments (HIM) in
conjunction with the k-Nearest Neighbor (k-NN) algorithm by using the Euclidean
distance. Third, when the object is classified, the MMP sends a signal through the serial
interface indicating which gate is to be opened. If the waste was not recognized, it is
deposited in the default container.

Several model designs for this project are shown in the figures attached.