It is in equal parts science and art. If the information is analogue, then one has to construct an appropriate converter – adapted to the nature of the signal and not introducing additional distortions. The result is influenced by everything: mathematical formulation of the problem – selection of appropriate formalism, selection of processing algorithm (digital filter or maybe neural network?) and implementation (microcontroller, FPGA element, DSP signal processor or a PC?). These are decisions on which the data processing depends as being successful, sufficiently fast and how they will influence the cost of the system as a whole.

Image processing is an equally complex matter. In particular, when the system has to make real-time decisions on the basis of imagery from several cameras. There are many proven algorithms and action schemes covering different aspects of extracting information from images. They comprise specific „building blocks” and are invaluable when building a processing application. Each issue is specific, which means it requires an individual approach. It may be high operation speed requiring special algorithm optimisation, it may be particular lighting which requires special selection of their working parameters or a non-standard shape of detected and classified objects, which requires creation of a new, specialized algorithm. This individual approach  to a problem is our motto.