Research On Laser Vision Tracking System For External Fillet Weld On Angle Steel Of Arm Frame On Crane

To realize automatic welding in the tower crane manufacture, a high reliable and durable laser vision based seam tracking system is proposed for welding the external corner joint between angle irons of arm frame.In this paper, the hardware design for the submerged-arc welding seam tracking equipment is based on an industrial computer, a CCD sensor, a step motor and its driver. The hardware uses the software developed by Borland Delphi to realize welding seam tracking.It is very difficult to detect clear and steady image owing to the noise interferences that exist in the nearby weld line caused by vast arc, spatter, etc. This paper corresponding filter system is applied to remove the light which wavelength out the range of 650±10 nm; And the remains is kept as light source. After affirms the relative geometrical position of CCD and laser by experiments, it can shot weld line by visual sensor, which contains abundant and objective information.The image is denoised through median filtering, and quasi-constant pixel number threshold or filtering three times with adaptive thresholds is applied to the image segmentation, and mathematical morphology is used to the binary image to filter isolated points and to detect edge, and the center line of seam is gotten through calculating the average of two edge pixels, with a particularly designed algorithm, characteristics and centre position information of groove can be calculated quickly and exactly.To the leading sensor, a tracking method through calculating seam track slope is put forward and PID control algorithm is applied in horizontal direction, and a fuzzy controller is designed in vertical direction.Practices show that the system can run steadily, and the tracking precision satisfies the need of welding quality completely. It is highly practical.