Fraunhofer Institute for Integrated Circuits IISSensor-based sorting has made enormous technological progress in recent years: data recorded using dual- and multienergy technology (DE- and ME-XRT) can be analyzed using machine learning methods. The procedure and the evaluation are very fast and at the same time very reliable. This makes them particularly suitable for harsh environments with high levels of dust and moisture.
© links: © Adobe Stock/luca piccini basile, rechts: Adobe Stock/ Kavik
Sensor-based sorting
We offer industrial X-ray sorting systems for recycling, mining and food production
We supply everything from a single source: from development, design and layout to the production of turnkey inline systems. The system can be integrated into the process through customer-specific control and individual adaptations, thus ensuring fully automatic evaluation and assessment of the material flow.
Dual-energy X-ray technology meets AI
With dual-energy X-ray technology, two separate images are aquired (low- and high-energy). Further, the measurements of the two images are aligned by using calibration methods. With precise knowledge of the X-ray system and using material specific attenuation, material properties can then be calculated from the measured gray values of the objects. The data can be processed further using artificial neural networks. This enables the differentiation between objects and thus their classification. Additional characteristics can be identified and extracted that could not be observed or were difficult to detect with classical computer vision. This makes the technology a multi-talent in the field of sorting and enables it to be used in numerous industries for a wide range of applications.
3D render AI artificial intelligence technology CPU central processor unit chipset on the printed circuit board for electronic and technology concept select focus shallow depth of field
Application example: Recycling aluminum with Dual-Energy
The aim is to sort the purest possible aluminum that can be used almost like »virgin material«. Fig. 1 shows that the objects are visually indistinguishable. In Fig. 2, the single energy image, the samples are also indistinguishable. We are able to detect impurities in the samples using Dual-Energy X-ray technology as shown in figure 3. Dark blue stands for high-purity aluminum, green for light impurities with silicon (6-10%) and yellow for impurities with heavier metals (copper, steel 0.5-2%).