Research on structured light vision seam tracking system based on RCNN
- Resource Type
- Conference
- Authors
- Liu, Yang; Wang, Tianqi
- Source
- 2022 IEEE International Conference on Advances in Electrical Engineering and Computer Applications (AEECA) Advances in Electrical Engineering and Computer Applications (AEECA), 2022 IEEE International Conference on. :1158-1161 Aug, 2022
- Subject
- Communication, Networking and Broadcast Technologies
Computing and Processing
Power, Energy and Industry Applications
Welding
Computational modeling
Training data
Predictive models
Laser modes
Prediction algorithms
Feature extraction
Structured light vision
Seam tracking
Convolutional neural network
Feature recognition
- Language
Aiming at the problem of locating seam feature points in seam tracking,a structured light stereo vision system based on line laser sensor is designed to realize real-time collection of images of weld lines. An algorithm for seam tracking based on region of convolutional neural network is proposed. First, randomly select two images from the collected seam laser fringe images to form an image pair, These image pairs form the training data set to train the model. Then, the search image that manually mark the location information of the seam feature point and the query image are input into the model for forward transmission. The final prediction output is the location of the seam feature point of the search image. A large number of test results show that the algorithm can accurately locate the location of the seam feature points in the seam laser stripe images.