Deep learning-based molten pool image geometric feature extraction method and system

Abstract

The invention provides a deep learning-based molten pool image geometric feature extraction method and system. The method comprises the steps of image acquisition, image preprocessing, image data set making, image segmentation network model construction and training, molten pool image online monitoring and molten pool geometric feature extraction and calculation. According to the invention, the morphology and evolution behavior of the molten pool in the laser welding process can be clearly observed, and interference information is less; according to the invention, a U-Net image segmentation network is used for extracting a molten pool area for laser welding for the first time, an improved eight-direction Sobel edge detection algorithm is adopted for extracting edge information, and the extracted edge information is more accurate; the molten pool area image extraction algorithm based on deep learning has strong anti-interference capability; compared with a traditional image processing algorithm, interference of splashing and plume can be better eliminated from the image, the anti-interference capability is good, and robustness is strong.

Description

technical field [0001] The invention relates to the field of laser welding technology and deep learning, in particular to a method and system for extracting geometric features of molten pool images based on deep learning. Through the image segmentation algorithm in the field of deep learning, the information of the laser welding molten pool area is accurately extracted, and the geometric features of the molten pool are extracted based on this. Background technique [0002] As one of the important application fields of laser processing technology, laser welding is a method of welding with the heat generated by bombarding the welding area with high-energy laser beams as energy. It has the advantages of less heat input, high welding precision, good quality, thermal deformation of joints and Due to the advantages of small heat-affected zone, it is widely used in aerospace, metallurgical machinery and other fields. In laser welding, the keyhole and its surrounding molten metal p...

Images