Abstract:
Image processing is a one of the popular techniques used in industrial quality
controlling. Such a system usually consists of image acquisition and processing,
followed by actuation operations. Image processing provides several benefits such
as high accuracy, high throughput rates, reduced labour requirement and reduced
fatigue for manual labour. Since it is naturally a non-contact measurement process,
it has wider range of applications. In this background, this research aimed to
develop a sorting conveyor system based on image processing in order to sort object
of different colour, shape, and barcode. A custom-made image processing software
was developed in C++ using the OpenCV library. Output signals from the image
processing software were used to drive a microcontroller which operates the
conveyor system. The conveyor system consists of a belt conveyor driven by an
electric motor and a rotary indexing table installed at the end of the conveyor
having radial compartments to store items. When items are sent through the belt
conveyor, based on the type of objects identified by the image processing software,
the indexing table is indexed to pre-set positions such that the respective types of
items are dropped on to the relevant compartment. Based on the developed
prototype, a series of experiments were conducted with the objective of fine-tuning
the software parameters to improve the identification of objects. Eventually it was
found that the system can sort items at a maximum frequency of 1 Hz, based on
colour, shape or barcode. Additionally, the software was further developed to
count items, which can effectively work even at 10 Hz. The proposed automated
system provides a cost effective, but flexible means of sorting items using image
processing. With further developments, this system can effectively be used for
industrial sorting applications.
