A clean image
Recent developments in optical technologies are helping the recycling industry protect the environment.
Recycling plants are starting to use image processing systems that can recognise and separate waste materials faster and in greater volumes than ever before. Identifying and sorting domestic, office and industrial materials more reliably and automatically cuts both the financial and the environmental costs of processing these materials.
The spectral imaging system
Spectral imaging is generally used for sorting paper, cardboard and plastics. It makes the chemical, chromatic and geometric properties of objects visible so they can be classified. This new technology characterises the surface of objects by combining the advantages of spectroscopy with those of digital imaging. The system illuminates the waste material and both near-infrared (NIR) and visible light cameras look at which parts of the light is reflected, which will vary according to the material.
Unlike conventional imaging, where a camera determines a grey or colour value (e.g. RGB) for each pixel, spectral imaging systems make all the spectral information at a certain wavelength available for every pixel. Since the NIR light contains spectral information that is specific to each different type of material, this provides a 'fingerprint' that enables it to be reliably identified. The image therefore can contain valuable information on the molecular or atomic composition and structure of the materials. It corresponds to a three-dimensional matrix with two spatial and one spectral dimension - information that would be unobtainable with other technologies.
Another key benefit is that the NIR emissions are not absorbed by surface coatings on the material, such as dirt or ink from labelling and printing.
The visible sensor determines the different colours of transparent and opaque objects (such as PET and PE bottles), sizes, shapes and positions. Adding a CMYK sensor detects and identifies paper/printed cartons that have been printed with CMYK technology.
Spectral imaging is incorporated in the Spectrosort system, which is used by a number of well-known recycling enterprises in Germany and Austria. On average (with two systems each) they sort up to 50,000t of waste paper per year. Each line can sort up to ten tonnes per hour, depending on the width of the conveyor belt.
The Spectrosort system comprises a light source, a scanning area with the sensors, a control panel and an array of air jets. Fast scanning sensors analyse the waste coming in on a conveyor belt to rapidly identify the materials, shapes, textures, colours and positions of objects. The air jets can then blow particular material types onto a second conveyor belt, while the remaining material goes on to a third belt for further sorting or disposal. The system can sort objects as small as 12mm. Investors in the technology can raise their capacity or reduce staff at their post-sorting stations, but at the same time improve the quality of their sorting.
Spectral imaging technology is not limited to the recycling industry. It could open up a wide range of applications for scientists and engineers in medicine, agriculture, the processing industry, recycling, chemistry and forensic medicine.
The Spectrosort system was originally developed by Carinthian Tech Research (CTR) in Austria. CTR's Spectrosort T business, including both the sensor technology itself and all the intellectual property rights on recycling applications, were recently acquired by Norway-based TiTech Visionsort.
CTR develops high-tech solutions and services in the field of optical sensors, image processing and microsystem technology and creates high-tech industrial applications for measuring, analysis and automation applications. It works closely with the Fraunhofer Institute.
Image processing components, such as the camera, the objective lenses, the image acquisition card and necessary cables for the Spectrosort project were supplied by the Stemming Imaging Group.