Vision systems

The progressive miniaturization of manufactured devices, increasing quality requirements – both from a functional and aesthetic point of view – and the need to optimize production costs, make conventional methods of quality control on the production line turn out to be insufficient. The most difficult tasks can be met only by solutions tailored to the 21st century – including automatic systems vision , which belong to the key areas of modern robotics and industrial automation. One of the most important applications of machine vision is solutions for automotive , but image processing technologies are also widely used in automatic assembly of electronics, the medical industry or production household appliances.

Vision systems are used to control the quality of production on automatic lines, where high repeatability of components is required. They are recommended where the line efficiency and the complexity of details is so high that the operator is unable to detect all errors.

Main use of vision systems:

 

Quality / presence control:

  • automatic application paths, e.g. adhesives,
  • presence of elements,
  • the correctness and completeness of component assembly,
  • component positions.

Replacing many sensors to verify the configuration of a transport pallet being converted depending on the reference – automatic assembly systems in the automotive industry.

 

Automatic identification:

  • reading 1D and 2D codes,
  • reading alphanumeric signs,
  • element encoding read.

Positioning:

  • positioning of robotic systems,
  • path correction,
  • item orientation before downloading.

Vision sensors – simple tasks, repetitive components, big difference between OK and NOK components:

  • component presence control (is / is not available),
  • component type control (shape / color / surface),
  • configuration control (appropriate markers in given areas).

Vision systems – smart camera (advanced tasks):

  • positioning, in a system calibrated to real units,
  • OCR / OCV reading and component type identification, production tracking, product sorting,
  • path control (continuous, open, complex shapes),
  • minimizing the influence of external conditions through the use of image filters and advanced possibilities of dynamic parameter correction for given image areas.

PC based vision systems – complex tasks:

  • advanced multi-camera measurements,
  • very high line performance,
  • complex control algorithms,
  • 3D control.

 

 

Advantages of using vision systems:

  • component quality control automation,
  • possibility of automatic correction of the position of manipulators and robots,
  • automatic reading of product markings (expiry date, batch number) – elimination of defective products
  • easy integration with control systems (digital I / O, RS-232, Ethernet RT),
  • high work stability – the ability to filter the image and dynamically select parameters even with variable parameters of the controlled product,
  • built-in automatic control program test mode,
  • the ability to register control photos during system operation, registration of NOK measurements for subsequent analysis and parameter adjustments,
  • control of many parameters with efficiency up to 750 measurements / min.
  • programmable user interface – the ability to define access levels based on user identification parameters from the vision system level or the ID from the control system.

 

Vision systems, or … machine vision in practice

The operation of modern video systems is based on several basic steps. The first one is image acquisition, i.e. taking an image of a scene (a workpiece or a group of objects) from the camera and sending it to a specialized computer that deals with further processing of the data received on a regular basis. A frequently used intermediate device is the so-called frame grabber, which allows fast-flowing video sequences to be buffered before the computer receives them. The software, based on efficient digital image processing algorithms, then performs a series of operations, starting with pre-processing, which can be imagined as photo retouching in a popular graphics program. Changing brightness, contrast, filtering, scaling, rotating or edge detection are just the simplest examples of such operations.

 

The next steps, performed on the initially corrected photo or video, depend on the expected functionality. Most often, image processing requires the identification of individual objects (according to shape and sometimes also color) and comparing them with a pattern saved in the computer’s memory. In this way, the software is able to detect irregularities, i.e. any deviations from the known pattern – thanks to this, it becomes possible to identify parts incorrectly mounted (e.g. in the wrong orientation in relation to the part base), missing or deformed parts. However, because it is impossible to meet all the requirements set by recipients with one universal software, in each case our engineers develop dedicated systems vision tailored to the client’s needs.

 

Vision capabilities

Machine vision systems are perfect for assembly validation applications, whether manual or automatic, upstream of the assembly line. A good example is the analysis of the correctness of the automatic application of adhesives or seals, thanks to which it is possible to detect details with a missing, incomplete or deformed application path. In addition, the vision systems can scan 1D and 2D codes, decode alphanumeric and other element coding systems, using this information independently or in combination with image analysis of geometry and / or color.

 

The rich capabilities of the vision software provide support for the positioning processes of industrial robots and machines. These systems also perform orientation of elements before taking for further processing or assembly, also with the use of 2D and 3D (stereovision) measurements carried out with the use of precisely calibrated sets of cameras. Easy and reliable low-level integration (e.g. via digital I / O lines or serial links with PLCs) or high-level integration (with SCADA systems, e.g. via Ethernet RT) enables the integration of data from the vision system into the production monitoring process. The most advanced applications include, among others solutions for automotive , although the possibilities of vision systems give invaluable services also in all other industries. At DS-Technic, we offer comprehensive support from the acquisition of customer needs, through conceptual development, to the implementation of a ready-made solution, installation and integration with an existing or newly created industrial automation system.