Measurement of pipe and profile diameter using camera and image processing

This thesis is coded in Python and features an algorithm that performs real-time image processing to measure the outer diameter of pipes and profiles using a camera. In the iron and steel industry that produces pipes and profiles, continuous diameter measurement is required. This measurement is necessary because different diameters are produced in various locations. During production, due to the positions of the rollers on the slabs, diameter variations may occur. Measurements are currently done manually using calipers, which can result in the production of faulty pipes and profiles. These products cannot be sold and are categorized as waste, leading to time loss and significant financial damage. The developed algorithm utilizes image processing methods via the OpenCV library. Its purpose is to continuously measure diameters from the camera, ensuring correct production on the first attempt. Since the system automatically performs diameter measurements, this approach eliminates the need for manual measurement. Separate options for pipes and profiles are presented on the screen, allowing measurement for both. Outside the tolerance values entered, warnings are provided in red, indicating whether the measurement is too large or too small. Calibration must be done according to pre-measured pipes or profiles based on the entered values. The distance and angle of the measured pipe or profile from the camera need to be adjusted to achieve calibration. Contours and thresholding are done according to the grayscale, increasing measurement accuracy by enclosing the gray color within a frame. The software utilizes an external camera as equipment, facilitating image capture within the desired positional range. The camera distance varies according to the diameter measured by the measuring device.