Sensor-Agnostic Material Classification Scheme – An Approach Derived from an Infrared Material Dataset Supporting Virtual Perception Testing

A material classification scheme is derived by structuring acquired data of a proprietary measurement setup based on a time-of-flight (TOF) camera operating in the near-infrared (NIR) spectrum, at a wavelength of 945 nm. This setup takes surface reflectance measurements representative for lidar sensors. Our approach aims for a sensor-agnostic material classification method with several levels of detail (LOD) based on mutually exclusive material classes to support and simplify direct material assignment. A major goal represents extensibility, general applicability, and facility of many different applications. The developed scheme may benefit different potential user groups, which range from measurement technicians, who are concerned with taking a multitude of different measurements to extend datasets in straightforward manner, to virtual artists, who are provided with a way of relating appropriate entries of the scheme to physical properties.