Laser welding quality monitoring method and system based on CNN and optical signal feature extraction

Abstract

The invention discloses a laser welding quality monitoring method and system based on CNN and optical signal feature extraction, and the method comprises the steps: decomposing a laser welding photoelectric signal through an adaptive noise complete set empirical mode decomposition method, and obtaining a photoelectric signal IMF component; performing feature parameter extraction and selection processing on the IMF component of the photoelectric signal to obtain effective feature parameters of the photoelectric signal; a one-dimensional convolutional neural network model and a two-dimensional convolutional neural network model are combined to construct a laser welding quality detection classification network model; and performing laser welding quality monitoring on the effective characteristic parameters of the photoelectric signals based on the laser welding quality detection classification network model to obtain a laser welding quality penetration state monitoring result. By using the method, the effective time-frequency domain characteristics can be obtained, so that the welding quality state can be quickly and accurately identified. The laser welding quality monitoring method and system based on CNN and optical signal feature extraction can be widely applied to the technical field of laser welding optical signal monitoring.

Images