Research on acoustic monitoring method of laser paint removal based on Bayesian discriminantion

In order to solve the problem that acoustic monitoring method of laser cleaning is difficult to be practically applied, the Bayesian discriminant method was used to conduct theoretical analysis and experimental verification. The paint removal process was divided into three categories: cleaning, cleaning completed and no damage to the substrate, and damage to the substrate. Combining the photoacoustic effect, the change of the paint removal sound signal during the cleaning process was analyzed, the characteristic parameters were extraced to establish a discriminant model. The accuracy of training samples reaches 99%, and the accuracy of test samples reaches 98.7%. The results show that the method has high accuracy and practicability. It can provide a reference for the research of laser cleaning acoustic monitoring.