Abstract: In the process of bad sorting detection of pitch diameter of cold oil joint thread, the detection accuracy and efficiency are low due to manual detection. Aiming at the problem, a sorting detection system of bad pitch diameter of oil pipe part thread based on machine vision was proposed. Firstly, in the pitch diameter detection model of machine vision detection, based on the principle of three needle pitch diameter measurement, the virtual measuring needle positioning algorithm of four needle method based on machine vision was innovatively put forward. The algorithm obtained the straight line of dental floss through Hough transform, combined the method of tangent circle positioning to automatically obtain the pitch diameter of the thread, and realized the automatic measurement of pitch diameter. Then, the machine vision system was selected based on the evaluation standard of the cold oil pipe thread, the overall design of the cold oil pipe thread diameter detection system based on machine vision was completed, and the control program process of the machine vision was completed too. Finally, the system detection accuracy and repetition accuracy were measured. The experimental results show that the system can screen the pitch diameter of the thread of the product, the detection accuracy is 0.01mm and the repetitive accuracy is 0.001mm, which provides a new solution and special machine design for the quality inspection of the same type of products.

Key words: hydraulic piping system; pipe joint; thread's pitch diameter measurement; three-point method; thread measurement algorithm; detection accuracy; Hough transform