METHOD AND APPARATUS FOR TRAINING PREDICTION MODEL BETWEEN LASER PROCESSING PARAMETERS AND REFLECTION SPECTRUM

摘要

Disclosed are a method and apparatus for training a prediction model between laser processing parameters and a reflection spectrum, wherein multiple reflection spectrum distribution curves are taken as input features of an input layer of a neural network, and multiple corresponding laser processing parameters are also taken as output features of the input layer of the neural network to train the neural network, so as to obtain a prediction model between the laser processing parameters and a reflection spectrum of a surface of a material. In addition, a method and apparatus for using a neural network prediction model to predict laser processing parameters are further comprised. By means of the method and apparatus, a prediction model, obtained by means of training, between laser processing parameters and reflection spectrum distribution curves can rapidly realize laser coloring processing with complex color distribution on a surface of a material, thereby accelerating the industrialization progress of femtosecond laser processing technology.