Computer vision-based industrial automation technology dates back to the late 1960s, but has only recently gained practical application and popularity in industrial enterprises. It is a branch of computer technology and a scientific field, the motivation of which is to replace the human ability to perceive the image with the eyes, to understand it and to interpret it. Approximately 80% of all information is perceived by the human eye and machine vision is able to obtain this and some other measurable information by technical means.
In terms of measuring methods, machine vision is one of the non-contact methods, and thanks to this, machine vision systems can be integrated into existing production plants and equipment without major modifications. Another advantage is the flexibility, which allows the system to be reconfigured, often simply by modifying the evaluation software. This requirement may arise, for example, when an extension of the production program, a change in a process, or the need for further evaluation.
Machine vision finds its application mainly in the light manufacturing industries, where it is used mainly for control and management of the production process. Its main task is to ensure the required quality or to monitor quantity.
The simplest inspection systems can only provide basic information on the status of the inspection, such as marking the product as perfect or scrap. More complex systems then alert the machine operator to an error or, with the help of simple automation, can eliminate unsatisfactory products from their own production. Sophisticated systems communicating with other devices, with which they pass control commands or extensive data structures, are then developed for complex evaluation tasks. The most common use cases, including specific examples, may be the following:
- production volume measurement – counting of bulk products on the conveyor and workplace;
- completeness check – detection of package integrity or control of package contents;
- color identification – control of printing or coating, evaluation of color of moldings;
- recognition, positioning and sorting – assembling the right part in the right position;
- reading and verification of 1D and 2D codes and reading of text (OCR) – advanced inspection during assembly;
- surface inspection – torn edges, scratches, cracks, undercuts, burrs;
- measurement and control of tolerances in 2D and 3D – position, size, hole spacing, hole size;
- robot guidance – product handling operations.
Benefits of machine vision
According to DZOptics, machine vision systems contribute to achieving a high degree of competitiveness. At the same time, these systems, compared to the use of human production operators, also have many other advantages.
The main advantages of machine vision include:
- reliability – machine vision systems are not missing compared to production operators;
- repeatability – production operators do not always work the same, unlike automation by machine vision;
- capacity – thanks to machine vision, high production efficiency can be achieved;
- operating costs – operating costs of a machine vision system are lower than manpower;
- visualization – outputs from the system can be visualized into evaluated images.
In addition to the technical aspect, the economic one is no less important when considering the use of machine vision. An integral part of any such project in which the system is deployed is a return on investment analysis. Machine vision automation is often approached in order to increase quality and reduce production scrap. They are factors that usually significantly increase production costs. However, worse cases can be repeated complaints from customers, which sometimes result in the termination of business relationships. These inconveniences can be prevented by using machine vision. In the long run, it is beneficial to build a good business name and strengthen its market position.