Extracting Raw Materials from Bulky Waste

ASKIVIT: Waste Wood Recovery from Bulky Waste using Artificial Intelligence and Image Processing in the VIS, IR and Terahertz Range

It conserves resources and improves profitability – recycling valuable raw materials from accumulated bulky waste is beneficial in many ways. Sorting bulky waste by hand creates employment opportunities, but is cost-intensive and thus has a negative impact on the actual recycling rate.

Automated processes based on various image capture and image processing methods as well as artificial intelligence (AI) make efficient sorting of bulky waste possible. In the development of such processes, we are initially focusing on wood and wood-based materials in the »ASKIVIT« project. In addition, we are testing whether the process can also be used for non-ferrous metals in parallel, because in contrast to magnetic metals, non-ferrous metals are difficult to separate from bulky waste. We as Fraunhofer ITWM contribute our expertise in terahertz imaging to this process.

With Sample Waste and Different Techniques to the Result

A representative, artificial bulky waste assortment as well as corresponding sample parts for the project are supplied by the Fraunhofer Institute for Wood Research, Wilhelm-Klauditz-Institut WKI. Thanks to two industrial partners from the waste management sector, real bulky waste samples are available.

Our scientists use these samples to perform measurements with terahertz imaging techniques. The shredded bulky waste is transported along a conveyor belt and examined from above using various techniques. These techniques are specifically adapted to the bulky waste sorting environment. In addition to terahertz imaging, other image acquisition and processing techniques are used:

  • Conventional image acquisition technology in the visible spectral range (Fraunhofer Institute of Optronics, System Technologies and Image Exploitation IOSB)
  • Near-infrared spectroscopy (Fraunhofer IOSB and Fraunhofer WKI)
  • Active heat flow thermography (Fraunhofer WKI)


More Visibility Through Terahertz Imaging

Often the materials to be recycled, such as wood and metal, are hidden by other materials, such as upholstery. Techniques that only look at and evaluate the visible surface do not detect these hidden raw materials. This is where terahertz technology comes into action: with terahertz imaging techniques, it is possible to detect deeper objects through non-metallic coverings.

For the »ASKIVIT« project we use a terahertz sensor, which is constructed as a line sensor and works according to the MIMO concept (Multiple-Input-Multiple-Output). So far, only flat, well-defined objects have been investigated using this technique. Both the sensor geometry and the reconstruction algorithms have to be adapted to the irregular surface of the bulky waste. While at the beginning of the project the terahertz sensor will be tested on well-defined samples, towards the end of the project a field test will be performed in a sorting plant using all techniques.

The measurements from the acquired sensor data are then processed, fused, and characterized by the KIT Institute for Industrial Information Technology IIIT. Artificial intelligence methods – especially deep artificial neural networks (ANN) – are used for this purpose. The ANNs are trained using sample data from the IIIT as well as bulky waste samples on a demonstrator at the Fraunhofer WKI. Fraunhofer IOSB is conducting field tests on the practical and market suitability of the overall system in a sorting operation.

Rapid Volume Inspection
© Fraunhofer ITWM
Fast inspection through the MIMO Terahertz System.

Relevance of an Intelligent System

More than two million tons of bulky waste are generated in Germany every year. Depending on the regional disposal concept, this consists of up to 50 percent wood. Recycling – instead of incineration or landfilling – raw materials such as wood and non-ferrous metals makes particular sense from an ecological point of view. Due to the increasing demand for wood and the simultaneous efforts to further minimize or even eliminate the use of the forest, the exploitation of waste wood is becoming more and more relevant.

Trained employees in disposal companies recognize wood-containing parts from pre-crushed bulky waste, but this selection is not done without errors and is very strenuous.

Our vision behind ASKIVIT:
an intelligent system that sorts bulky waste accurately and without fatigue – even without the prior shredding.

Economic Benefits

Advantages of an automated process can also be seen from an economic point of view:

  • Disposal companies benefit from the cost-efficient sorting as well as from the increased amount of recovered raw materials, which can be sold again.
  • The wood-based materials industry is becoming less dependent on fresh wood.
  • The broad raw material base and increased efficiency of recovering waste wood from bulky waste creates economic benefits for companies that manufacture, process or use materials.
  • Consumers themselves could also benefit if bulky waste becomes a valuable raw material. However, it is a political question whether the disposal of bulky waste should be free of charge or even profitable for consumers. 

Our Project Partners

  • Fraunhofer Institute of Optronics, System Technologies and Image Exploitation IOSB (project coordination)
  • Fraunhofer Institute for Wood Research Wilhelm-Klauditz-Institut WKI
  • Institute for Industrial Information Technology (IIIT) at the Karlsruhe Institute of Technology KIT
  • ALBA Braunschweig GmbH
  • DIE GRÜNEN ENGEL Disposal and Logistics GmbH
  • Dieffenbacher GmbH Mechanical and Plant Engineering

Project Duration and Funding

Duration: 01.07.2021 until 30.06.2024

The project is funded by the German Federal Ministry of Food and Agriculture (BMEL). The project executing agency is the Agency for Renewable Resources (Fachagentur Nachwachsende Rohstoffe e.V.).

Official detailed project title: Waste wood recovery from bulky waste by artificial intelligence and image processing in the VIS, IR and terahertz range. Subproject Two (ASKIVIT-Thermo) entitled: Experiments on the detection of wood and wood-based materials in bulky waste using active heat flux thermography.