Computer Vision

In image processing measurement is based on light. The quality of the system thus depends considerably on the lighting conditions and their continuity. The human eye adapts unnoticably to new color or lighting conditions. Perspective distortions, indistinct one and noise hardly represent a problem for the brain. In the image processing it is an art to arrange the stable recognition in precense such influences.

Input and Output

The process of image processing begins with the picture as input. The picture can be taken from different sources. In Vision2u pictures can be *.bmp; *.png, *.gif formats and videos are loaded from disk. Also the use of camera, scanner or Screenshots can possibly be a source of picture.
Beside the sources of picture different data sources can also be used : User inputs, text files, Web services, file system, date, time.


With the help of signal processing algorithms, we can filter noise and sharpness from pictures. Also the perspective distortion of objectives (ton/cushion distortion) can be corrected with filters.


If a picture is only once prepared, then it is not a problem to recognize distances between edges or objects or to count the quantity from and to objects. Before the object recognition it must be ensured that the correct parts are found. Then they can be counted and measured industriously. For measuring lengths and distances a calibration of the image input parameter or later a conversion is necessary to get a desired unit.


A frequent task in the image processing is supervision. For this a comparison of the current picture with a reference or the predecessor picture is accomplished. Differences can be seized and differences starting from a threshold value be declared by the comparison.

Computer-Assisted Reading

With Vision2u reading of Bar code (EAN-13), aileron code or writing recognition by means of ocr are possible. Naturally before processing bar codes or texts are converted into a picture.


For the evaluation of picture information, the picture must be first analyzed. The picture can be compared for this purpose by reference colors or brightness. Alternatively a segmenting of the picture can take place. For this the picture is divided into a given number of segments. The following methods of analysis are available:

Object Searching

First the position of objects must be determined in the picture for the image processing. For this there are the functions of the pattern recognition. A reference picture or a description of picture in the picture is looked for. The following search functions exist:

Image Editing

In addition to the processing of pictures their manipulation (treatment of pictures) is possible. For this the following functions are available: Cut out, scaling, rotating, color change, change of brightness, picture operations, lines, texts, rectangles, circles.


The best image processing is useless, if thereafter no output is generated. Under Vison2u different image outputs are available:

Additionally results of analysis can be passed over to the following interfaces: