Intelligent Process Parameter Tuning of Laser Cleaning Based on Image Feature Analyses

An image feature analysis-based method is proposed to conquer the intelligent process parameter tuning problem of laser cleaning machine. First, lots of metal corrosion images and their corresponding laser cleaning process parameters are collected. The settings of the laser cleaning process parameters come from the experts’ experiences. Second, some image features, i.e., the colour and the texture features are computed. Third, the Back Propagation (BP) network is employed to build the connection between image features and process parameters. Finally, when this system works, the image features of workpiece will be calculated, then the BP network will estimate the proper process parameters for laser. Many experiment results have verified the correctness of proposed method.