It is probably one of the most demanding tasks of industrial automation technology: imitating the functions of the human eye and equipping machines with corresponding image processing capabilities. Below you will find an overview of challenges and technical progress.
Replacing the human eye and brain by technology – this might sound absurd to laymen. Among specialists, however, this approach is seen as a groundbreaking development that has gained significant momentum in the past few years. Machines with the ability to ‘see’ can take over more and more tasks in the industry, draw conclusions from scenes observed and thereby support humans. In order to understand the industrial imaging technology behind, one has to understand the mechanics of human image processing. The basis for this is the human eye, which perceives its environment and transfers corresponding information to our brain. Here the information is interpreted and our brain initiates corresponding reactions. While reading this article, you perceive individual letters, your brain establishes a context and you might even judge what you have read. Or, to name another example: Employees of a factory can identify levels of tanks, silos or hoppers with their eyes and, if necessary, initiate the impulse to stop the machine.
Technology behind industrial imaging
In industrial imaging, tasks done by the human eye and brain are carried out by cameras or image capture sensors in combination with illumination units and processors with included image processing software. Both for the human and the technical process, it is not the object considered that is essential, but the light that the object reflects to the eye or to the camera or sensor. This principle is also used for ifm’s O3D sensors. Thanks to the patented PMD technology using a time-of-flight 3D chip to determine the distance and grey-scale value for each pixel, the sensors can be used for three-dimensional real-time detection of any scene. Application areas range from level measurement and point level detection to volume determination and completeness monitoring.
One of the biggest challenges currently faced by industrial imaging is to make the different components understand and interpret image data – something done largely intuitively by humans. For example, over time, we learn to interpret the facial expression of other people and to identify certain patterns as right or wrong. Imaging systems also identify objects correctly, if they have been programmed or taught accordingly beforehand. This also applies to ifm’s O3D: If the data determined by the sensor deviates from the defined target state or exceeds defined limits, the sensor switches, thereby preventing production faults or machine failures.
You can find more information about industrial imaging products from ifm – for example the O3D – here.
You would like to learn more about the process of digital image processing? Read the report “Machines with eyes wide open” by VDMA Machine Vision (Machine Vision department of the Verband Deutscher Maschinen- und Anlagenbau e.V.(German Engineering Federation)).
Picture credits:
