Industrial image processing - a key competence in industry 4.0
The term Industry 4.0 describes a future vision for the digitalization of industrial production. Mechanical industrial production is linked with modern information and communication technology in such a way that a largely self-organized production is created: people, machines, plants, logistics and products communicate directly with each other. The operating and production data collected in this way make it possible to optimize not only individual production steps but also entire value-added chains.
Industrial image processing plays a key role here, as it allows individual production steps to be visually documented in detail and analyzed in real time with the aid of image recognition software. The obtained data can be used for quality assurance, process control and process optimization.
The human machine fitter of an industrial image processing system contributes his cognitive skills and his expertise by determining the parameters to be recorded and defining the reaction of the system to the parameter values determined and storing these process data. In this way, the technical knowledge of the machine fitter is digitized, preserved and machine transferable.
The hard- and software of an industrial image processing system can read in the digitized process data and execute the process described in it at high speed, with constant precision and for a virtually unlimited time.
The described concept of industrial image processing is excellently suited if the image acquisition takes place under reproducible, constant environmental conditions (especially the lighting situation) and the measured parameter values trigger clearly defined reactions of the system.
It reaches its limits when environmental conditions can change or when interpretation of the determined parameter values requires human intuition, association or an (adaptive) discretionary scope. For in these cases, the human plant designer would have to be able to foresee every possible situation and determine the correct reaction of the plant for it. Where this is not possible, artificial intelligence methods can be used to train industrial image processing systems.