Machine vision for industry

Abstract

Nowadays, industry is changing its way of working in order to get more competitive. Industry wants to get more and more automated in order to reduce production times, increase productivity, improve quality production, at a cheaper cost, to be less wasteful, with less need to have a skilled operative, to be more flexible meaning easy to implement changes and possible to leave the automated process working with little supervision around the clock. Vision in Robotics helps controlling production in a relatively simple form, avoiding a skilled operative to spend his time ‘watching’ the machine doing his job. Moreover, the automatic inspection process does things faster and with improved quality than a human. Robotics Vision and Image Processing tools are the most desired tools for quality control in industry. With the use of one (or more) cameras, and a computer controlling and analysing the extracted images, a software tool can solve a problem in a relatively easy way. An initial investment is needed to buy all the necessary Vision Hardware, but software can be built by using a few existing tools. Instead of making a vision based program from scratch (re-inventing the wheel) to solve a specific problem, it is now possible to use existing image processing tools and build quickly and easily a software solution. These tools work on grey-scale image processing level. These high-level vision software tools do not require that the developer program at the pixel level, which makes the technology accessible even to users with little machine vision experience. To reduce the amount of image to analyse, the user can work on ‘regions of interest’, reducing though the time and space to analyse/store the image. A description of the most important tools are described and its basic principle of functioning is explained. These tools can then be integrated and work together in order to make the full solution.