Automatic Adjustment Method of Operating Parameters in Laser Cleaning Process Based on Image Processing

Abstract

When using laser cleaning on metals with different levels of corrosion, it is necessary to adjust the operating parameters of the laser to achieve the cleaning standards without damaging the metal itself. Traditional laser cleaning methods rely on experience, which are subjective and poorly adaptable. This paper proposes an automatic parameter adjustment method based on image processing. The method first extracts the corrosion image features of the metal and then establishes a parameter prediction model based on the Particle Swarm Optimization-Back Propagation(PSO-BP) neural network. The model takes the color and texture features of the metal to be cleaned as input and the operating parameters of laser cleaning as output. Comparative experiments show that the PSO-BP neural network prediction model has significant advantages over the traditional Back Propagation(BP) neural network model. It was also found that when using the combination of color and texture features of the corrosion image as the input to the model, the prediction accuracy was improved. Tests on 100 sets of experimental data show that the average prediction accuracy of the proposed method can reach over 85%, meeting the requirements of practical laser cleaning applications.

Keywords

• BP neural network
• laser cleaning
• image feature
• PSO-BP neural network