| Welding plays an important role in the production industry, which quality has a direct impact on production. The welding process real-time tracking has become an important means to ensure the quality of welding. There are much advantage for the visual inspection possess such as intuitive, fast and accurate. A seam tracking detection system is developed with the image processing techniques. By deeply researching on the characteristics of the weld image, a relatively perfect hardware detection structure is established which base on USB industrial camera. The software is developed with OpenCV function by Qt development tool under the system of Ubuntu. In this software system, the feature information of seam is extracted and the rapid detection of seam tracking is completed.Firstly, the general scheme of the detection system is designed. Then, this thesis selects the hardware, and analyzes the installation location. Based on these research results above, the hardware system is built.Secondly, the image processing technology is used to achieve the assessment of the vertical of camera, designs of camera distortion correction and calibration. In the image pre-processing, eliminate most of the weld interference image information by using a one-dimensional convolution templates. According to the characteristics of the laser light, a dynamic window procedure is designed, which reduces the data processing and improve the processing speed. In order to eliminate the interference information which is below the target threshold, a binary image segmentation is designed. Base on different forms of groove weld characteristics, subsequent image processing is carried out, and the relevant deviation information of weld is extracted.Finally, the software system of this paper is established on the Qt development platform. The detection precision, stability and rapidity of this system are checked through several experiments, and the results show that the system can meet the need of measuring. |
PDF Full Text Request