Online detection method and device for welding process information

The invention discloses an online detection method and device for welding process information, and the method comprises the steps: firstly, collecting the dynamic image information data of a welding process molten pool shot by an industrial CCD camera through a software development platform; molten pool image feature information is extracted through an image processing means; based on a deep learning algorithm, correlation between molten pool image feature information and a welding state is established; and performing state recognition on the molten pool image by using an algorithm model obtained by a deep learning algorithm, and finally realizing welding seam quality evaluation. In order to solve the problem that accurate description is difficult due to randomness of image information changes and diversity of welding states in the welding process, correlation information between images and the welding states can be learned through a neural network, and adjustment is conducted through a self-learning mechanis