Discovering new paths in machine vision - monitoring plasma coating process with hyperspectral imaging:
New possibilities for
advanced machine vision
Machine vision gives eyes to a computer by adding a specific number of video cameras and helping the computer to understand its surroundings. The resulting data is used in various applications such as automatic inspection, process control, and robot guidance.
With the possibility of using a hyperspectral snapshot camera this adds a new dimension to machine vision, since not longer three colors (RGB) can contribute to decision processes, but more than 100 spectral bands, covering a region even more than the eye could meet.
Machine learning enables
Furthermore, machine vision benefits from state-of-the-art analyses technologies. The magic word is machine learning, and although people today might easily think this solves every problem (which it unfortunately won’t), it tremendously helps processing big data in a short time.
Because this simply is what it is about in machine vision – retrieving and extracting the relevant optical information in order to trigger a subsequent process, in a limited range of time. In addition, not only can materials be distinguished, but even be qualified on their individual status. This enables a high number of possible applications in the industry.
Wide fields of applications in a quickly developing market
Through the newest technologies for machine vision like classification strategies based on machine learning the industry can benefit from hyperspectral imaging tremendously.
Spectral imaging will be the key differentiator with respect to quality inspection, for instance in the food industry: fruits, vegetables, meat, grains and even biofilms for food packaging.
The recycling market is also benefiting from this technology in terms of sorting different types of paper, plastics, and waste electronics, only to name a few.
Some error has occured.