Laser Vision Sensor Based Post Weld Inspection Technique
|School||Harbin Institute of Technology|
|Course||Materials Processing Engineering|
|Keywords||Laser Vision Sensor After welding detection Demarcate Image processing|
At present, the detection for weld appearance, mainly rely on the use of testing personnel gauge, ruler, magnifying glass tool manually to complete. The drawback is the low detection efficiency and accuracy, stability is not good, and not be able to achieve real-time on-line continuity check. After welding, laser vision sensing detection method can be a good solution to the above problem. Therefore, a preliminary study of the technology. Starting from the detection principle of laser vision sensing using laser tilt irradiation, camera a vertical receiving the program. Based on the size of the weld to be measured, the line length of the laser stripe and the CCD camera imaging relationship calculation and design of lasers and the positional relationship of the camera and a working distance, the detection accuracy of the theory of the design of the sensor reaches 0.08mm/pixel, description of the design of the sensor can achieve the accuracy of the measurements of the weld. In accordance with the four reference coordinate system, establish the mathematical model of a laser vision sensor, the analysis of the corresponding relationship of the laser plane on the object point and image point. Obtained on the basis of the mathematical model of the laser vision sensor requires calibration of the internal and external parameters. Were used the plane template method and stepped target method to complete the calibration of the camera internal parameters and external parameters of line structured light. Camera calibration method is simple and convenient, and easy to implement. The stepped target design overcomes the sawtooth-shaped target to obtain small quantities, the disadvantage of the feature point, the target material using the reflective properties of the metal workpiece to be tested, using the Levenberg-Marquardt optimization Solution to obtain a highly accurate external parameters. Analog signal of the laser-vision sensors capture images through image acquisition card signal into a digital signal sent to the computer by the computer image median filtering, thresholding, mathematical morphology algorithm processing to extract the laser stripe centerline. The results indicate that the extracted after processing of the above steps, a high accuracy of the centerline of the laser stripe, and good connectivity. Butt welds after welding detection technology, through to strike the inflection point of the strip center line, the slope and search boundary points identified include weld width, I weld toe angle, undercut, concave amount flat angle, the wrong side of size. The experimental verification shows that are better measurement accuracy and stability of the detection system of laser vision after welding. Finally, using Visual C programming language, system software development based on the modular thinking.