Machine vision applications

With the increased power of microprocessors, applications that can be solved successfully using vision systems are growing quickly. An application may require one or more image processing functions, which, when combined, create a solution. The detection range is wide and includes:

Shape or appearance. Compliance check

Current systems usually begin with a two-dimensional measurement operation to establish the displacement necessary for the object in an ideal position. Subsequently conducted operations compare the captured image with a standard object or geometric shape characterization. In this way, you can decide if the scanned object meets or not the predetermined requirements.

Detection of defects. Discrete elements

The detection of defects can be considered as a special case of control of conformity, in which objects do not offer specific forms characterizable. As examples, we quote the inspection of ceramic tiles in the pre-glazed for defects on the surface roughness, or the use of infra-red inspection of defects in glass bottles.

Detection of defects. Fabrics or laminated

The point of many systems of this type is to examine large areas at high speed, looking for relatively small defects. The material produced in rolls including steel, paper, photographic film, glass, plastic film and textiles. Some of these systems use laser scanning rather than linear cameras to form the image.

Color control

The color inspection is widely used in the pharmaceutical industry to confirm that the tablet or pill color has been placed in the correct packaging.

Dimensional or two-dimensional measures.

Dimensional measurements can be made by superimposing optical gages on the product image. This kind of measurement is very fast because there is no mechanical movement and exact positioning of the object to be measured is not critical (the system can determine the displacement X and Y of the object and the degree of rotation). Some systems offer sub-pixel interpolation, in appropriate conditions causes the system to be able to measure and, in particular, to detect changes or differences in measurement, rather than the thousandth decimal point offer other methods of metrology. The number of pixels in the image can be increased by using very high resolution cameras and obtain even more precise measurements.

Structured light triangulation and other techniques.

If a thin band of light is directed to a three-dimensional surface and is seen with a camera at an angle different from the illumination device, the apparent shape of the band can be used to infer the shape of the exposed surface along the length of the band. By scanning the band through the surface can generate a three-dimensional map of the observed object. It is often convenient to use a laser light source, since it can easily be refracted to form a plane of light.

Three-dimensional techniques.

You can also obtain three-dimensional information using automated stereoscopic vision.

Character recognition.

Character recognition can be based on simple correlation techniques (comparison), without specialized knowledge of how individual characters are formed, or you can use more advanced algorithms of artificial intelligence.
Vision systems can also be used to read bar codes with a higher degree of confidence that a laser scanner. Even if a large part of the bar code is obscured by dirt, a vision system, in general, you can read, whereas a bar code reader is not conventional.
In addition to bar codes, vision systems are also used to read two-dimensional matrix codes where codes may contain more information than one dimensional bar code.

Recognition of parts or components of the product

Recognition of parts or components can be used to identify which parts should be subjected to that process. For example, a vision system can distinguish between different parts of an item and instruct a robot to the correct assembly of the element.

Guided default

The default guidance is characterized by a situation in which an azimuthal camera takes a snapshot of the scene and the vision system directs a robot to pick up or leave an object in a particular position. Then the robot work in the dark. A typical application includes tasks palletising of heavy objects, such as engine blocks and crankshafts, but can also include much lighter work such as packing chocolates.

Continuous guidance

The continuous guidance involves a camera mounted on the arm or hand of the robot and the path of the robot is continuously corrected by the vision system. A common application is the guidance for arc welding, but the technique can also be used to control the way the application of sealants or adhesives for robots.