Software solution for Testing Image Processing Algorithm on Conveyor-Based Vision Systems

Abstract

Vision systems have been used in many applications that intends to reduce the need for human operators. This is especially true for tasks that are simple but repetitive in nature, which is largely applicable to most manufacturing and agriculture’s post-harvest processes. Many such processes utilize conveyor-based systems where the objects being processed are placed on a conveyor belt that runs through multiple processing stations. Implementing a vision system to capture images of an object that is moving usually requires setting up an imaging device to a working conveyor system. Getting a working conveyor system to be ready can take some time and consequently delay development work on the vision system itself, especially those involving image processing algorithms. This paper proposes a software solution that can be used to expedite initial work on such systems. The solution is written in C and is therefore easily ported to any development machine. A basic image processing library has also been developed so that it does not depend on any development library or suite, which is usually huge in size. Thus, the solution can easily be compiled and run on embedded development boards like Raspberry Pi - for a more portable solution.