Laser Vision Sensor Based Post Weld Inspection Technique

At present, seam surface inspection mostly relies on workers using measure gauge and magnifier. The method is inefficient, inaccurate and unstable and can not achieve real-time continuity measurement. Based on laser vision sensor, post weld inspection technique can solve this problem very well. So a pilot study was carried out in this technique.Depending on the measuring principle of laser vision sensor probe, the scheme of the sensor used oblique semiconductor laser and vertical camera. According to seam size, laser stripe length and imaging relation of CCD camera, the position and working distance between semiconductor laser and CCD camera was designed. The theoretical inspection precision of designed sensor was 0.08mm/pixel and indicated that designed sensor could attain the precision of seam inspection. In the basis of four reference coordinate systems, the model of laser vision sensor was founded and the relation between object point and image point was analyzed.We could got the inside and outside parameter from the math model of the sensor. The camera and structured light calibration was accomplished used plane method and step target method. The camera calibration method was simple, convenient and easy to carry out. The step target had overcome the defect that serrate target could not obtain more calibration characteristic points. Using metallic material that the reflect was the same as the seam to make the target and adopting Levenberg-Marquardt optimization algorithm, outside parameter was acquired.Via image card, the analog signal captured by Laser vision sensor was switched to digital signal presented to computer. Seam image was processed by the algorithm of median smoothing, threshold segmentation and morphology , then the central line of the laser stripe was extracted. The results showed that the extraction method has high precision and good connectivity.The post weld inspection used in butt weld characteristic information was studied. Seam macroscopic size including bead width, reinforcement height, toe angle, undercut, concavity, joint angle and mismatch has been identified. Test and verify indicated that post weld inspection system has high precision and good stabilization. Finally, used Visual C++ programming language, based on modularization, system software was designed.