Automated Vision System to Measure Weld Length of Hand Brake

To measure the quality of welding joints, various inspection methods like laser triangulation, X-ray diffraction, ultrasonic detection, and band-pass filters are practiced. Automated quality check methods provide increased quality production. In this research paper, to check the quality of welded joint on a handbrake lever and to make decision based on the acquired data, the part image is captured by a camera and it undergoes image processing methods in MATLAB. To highlight each weld boundary, region of interest is used. The highlighted regions are then processed by techniques such as segmentation, thresholding, dilation, erosion, and morphological masking. The following facts are observed that there is lack of clarity in the acquired image, improper threshold value set while segmenting the image, erroneous mode of shape masking, thereby generating gaps between pixels. Finally, the method holds good for this particular material with weld in particular coordinates. The presence and length of the welds present in the material are successfully obtained. This inspection system is restricted only for the weld length and weld presence of the handbrake lever application.