Automatic inspection of tire geometry with machine vision
Junfeng Li, Yuchun Huang · 2015
On-line automatic inspection system for measuring the tire geometry is proposed with the dedicated machine vision design. A sequence of the tire images are acquired while a tire object passes through the inspection system. The tire geometry, including the cutting edge, angle and width, represents the quality of the sequential tire objects that are connected together to make the inner tire belt. A sequential line matching algorithm is developed to measure the tire geometry automatically. With the input of the standard tire model geometry, Line Detection with a Priori (LDP) is conducted to locate the candidate lines of interest in a high speed. Laser light is cast to the tire surface from a small incident angle, which robustly refine the location of the tire edge in a image. Lines of the sequential images of a tire object are matched based on their orientation and distance in the Hough space. Lots of experiments in the Aeolus Tyre manufacturing Co. Ltd, China show that, the proposed system can inspect the tire geometry consistently with the human measurement and at the same time reduce the time of measuring to 20ms. It is promising for the industrial applications of high-speed measurement of accurate geometry with low cost constraints.