Challenges in automatic sorting of construction and demolition waste by hyperspectral imaging

Frank Hollstein; Íñigo Cacho; Sixto Arnaiz; Markus Wohllebe

doi:10.1117/12.2220769

Translator Disclaimer

9 May 2016 Challenges in automatic sorting of construction and demolition waste by hyperspectral imaging

Frank Hollstein, Íñigo Cacho, Sixto Arnaiz, Markus Wohllebe

Author Affiliations +

Proceedings Volume 9862, Advanced Environmental, Chemical, and Biological Sensing Technologies XIII; 98620J (2016) https://doi.org/10.1117/12.2220769
Event: SPIE Commercial + Scientific Sensing and Imaging, 2016, Baltimore, Maryland, United States

Abstract

EU-28 countries currently generate 460 Mt/year of construction and demolition waste (C&DW) and the generation rate is expected to reach around 570 Mt/year between 2025 and 2030. There is great potential for recycling C&DW materials since they are massively produced and content valuable resources. But new C&DW is more complex than existing one and there is a need for shifting from traditional recycling approaches to novel recycling solutions. One basic step to achieve this objective is an improvement in (automatic) sorting technology. Hyperspectral Imaging is a promising candidate to support the process. However, the industrial distribution of Hyperspectral Imaging in the C&DW recycling branch is currently insufficiently pronounced due to high investment costs, still insufficient robustness of optical sensor hardware in harsh ambient conditions and, because of the need of sensor fusion, not well-engineered special software methods to perform the (on line) sorting tasks. Thereby frame rates of over 300 Hz are needed for a successful sorting result. Currently the biggest challenges with regard to C&DW detection cover the need of overlapping VIS, NIR and SWIR hyperspectral images in time and space, in particular for selective recognition of contaminated particles. In the study on hand a new approach for hyperspectral imagers is presented by exploiting SWIR hyperspectral information in real time (with 300 Hz). The contribution describes both laboratory results with regard to optical detection of the most important C&DW material composites as well as a development path for an industrial implementation in automatic sorting and separation lines. The main focus is placed on the closure of the two recycling circuits "grey to grey" and "red to red" because of their outstanding potential for sustainability in conservation of construction resources.

Citation Download Citation

Frank Hollstein, Íñigo Cacho, Sixto Arnaiz, and Markus Wohllebe "Challenges in automatic sorting of construction and demolition waste by hyperspectral imaging", Proc. SPIE 9862, Advanced Environmental, Chemical, and Biological Sensing Technologies XIII, 98620J (9 May 2016); https://doi.org/10.1117/12.2220769

ACCESS THE FULL ARTICLE