Surface condition sensor, method for training an evaluation algorithm of a surface condition sensor and method for determining a surface condition parameter from a predetermined light signal

摘要

The present invention relates to a method for training an evaluation algorithm, designed as a machine learning method, of a surface condition sensor (1), which by the steps of defining (S1) a surface condition parameter, generating (S2) a light spectrum on the basis of the specified surface condition parameter and a predetermined model that includes the Modeled light spectrum that is reflected by a surface when the corresponding surface condition parameter is present, and the use (S3) of the specified surface condition parameter and the generated light spectrum is characterized as training data for the evaluation algorithm, with at least part of the generated light spectrum being used as a reflection signal. The present invention also relates to a method for determining a surface condition parameter from a predetermined light signal. The present invention also relates to a surface condition sensor (1) which has an evaluation unit (10) which uses the evaluation algorithm to determine surface condition parameters from the reflection signal (4).