Measurement and Inspection of Roundness Using Computer Vision.

Gadelmawla, E.; Khalifa, W. M.; Elewa, I. M.

doi:10.21608/bfemu.2020.127536

Measurement and Inspection of Roundness Using Computer Vision.

Document Type : Research Studies

Authors

¹ Production Engineering and Mechanical Design., Faculty of Engineering., El-Mansoura University., Mansoura35516., Egypt

² Production Engineering and Mechanical Design Department., Faculty of Engineering. Mansoura University., Mansoura 35516., Egypt.

³ Production Engineering and Mechanical Design Department. Faculty of Engineering. Mansoura University., Mansoura 35516,. Egypt.

10.21608/bfemu.2020.127536

Abstract

A common problem of quality control and inspection of circular parts is the measurement of their roundness. Errors in rounded objects can greatly affect the accuracy and function of assemblies. Therefore, there is a requirement to develop' automatic inspection methods for the inspection of roundness. This paper introduces a new approach for measurement and
inspection of roundness using computer vision. Tile vision system has been established in the metrology lab and it is used to capture images for objects 10 be measured or inspected. A software (named RoundVision) has been especially developed in-house using Microsoft Visual C++ to analyze the captured images and perform the measurement and inspection processes. Unlike the conventional methods. the proposed system has the advantage of being a non contact method and much near to the reality because all points on the contour of the objects to be measured are considered. In addition, the system is capable of measuring and inspecting multiple rounded features ·in the same part, whether it represents the external frame or internal holes. The system has been calibrated for both metric and pixel units. To verify the introduced system, sample circular parts were used to measure their roundness error by the introduced system and a commercial system (Round test RA-100), then the results were compared. The maximum difference between the results of the two systems was ± 7 .22%.

Keywords