The sorter uses hyperspectral imaging (HSI) instead of near-infrared technology. The HSI camera records high-resolution images and provides detailed information on material characteristics. All sensor data – including NIR, colour, 3D and metal detection – is processed via a sensor-fusion concept. The raw data is evaluated directly without additional interfaces, enabling loss-free data transfer to the detection algorithm, which is trained on extensive datasets.
Material allocation is displayed in real time. In lightweight packaging streams, both material types and object categories can be visualised. Additional sensors, such as 3D or metal detectors, can be integrated at raw-data level. Provisions are also made for future marker-based recognition technologies.
The system can be expanded with object-based AI when a colour camera is installed. This enables precise identification and targeted separation of predefined objects, supporting adaptation to changing sorting specifications.
The BRT Hartner AS is designed for materials that are difficult to separate using conventional processes. Its AI evaluates all sensor data to classify materials and objects, allowing differentiation of various plastic types and packaging formats. This supports sorting steps required for mechanical and chemical recycling, as well as the separation of high-calorific fractions in substitute fuel production. The system can also be used for general impurity removal.
Standardised, commercially available light sources are used, avoiding specialised components. According to the manufacturer, the system requires fewer light sources than typical HSI-based solutions, reducing operating and procurement costs.
Maintenance access has been integrated into the design. An extendable internal platform facilitates work on the separation unit, and the sensor unit can be optionally configured for easy cleaning and maintenance from a walkway.
