When the robot performs real-time pipeline weld tracking, the weld image obtained by the camera will be interfered with by solid arc noise and spatter, which makes it challenging to ensure the stability of the weld quality of the pipeline. This paper proposes an automatic weld seam feature recognition algorithm based on an improved U-Net neural network. The method extracts the global features of the weld image through the backbone network after down-sampling and upsampling in the U-Net network, fuses the laser stripe information at multiple scales, and utilizes the feature enhancement module to obtain more explicit weld feature images. Experiments have shown that the accuracy can reach 99.17%.
KEYWORDS: Convolution, Detection and tracking algorithms, Deep learning, Object detection, Education and training, Convolutional neural networks, Computing systems, Target recognition, Roads, Process control
Welding technology is one of the key processing technologies in the field of mechanical manufacturing and engineering construction. With the application of artificial intelligence in welding equipment and process control technology, the degree of automation, control accuracy and quality stability of welding technology have been improved to a certain extent. However, there is a lack of effective supervision measures for pipeline welding quality to ensure the smooth progress of process processing. In this paper, the improved YOLOv8 algorithm is used to identify and weld, and the original loss function is replaced by the SIoU loss function, which enhances the iterative speed and detection progress of the model. In addition, the depth separable convolution is used to replace the original void convolution, which makes up for the problem of information loss when the size object may exist, greatly improves the recognition ability and accuracy of the weld quality of the model, and lays a foundation for the improvement of the welding ability of the welding robot.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.