Abstract

LOPEZ, Carlos A.; FERRIN, Carlos D.  and  CASTILLO, Luis F.. An Approach for Applying Computer Vision in Refrigerator Quality Control on Assembly Lines. Inf. tecnol. [online]. 2019, vol.30, n.2, pp.143-156. ISSN 0718-0764.  http://dx.doi.org/10.4067/S0718-07642019000200143.

This paper proposes a computer vision methodology to detect internal accessories in real time and discern their correct assembly inside the refrigerators. Refrigerator returns due to aesthetic failures are one of the biggest problems facing the white goods manufacturing industry. Main cause of this problem is that aesthetic evaluation is a subjective task that depends on the visual and cognitive skills of quality inspectors. The methodology consists of an acquisition block of multiple views using RGB-D-NIR sensors and algorithms such as segmentation, object tracking, regions of interest detection and key point detection based veredict. Experimental results showed that the proposed methodology presents good performance and speed; besides, it does not introduce cadence to the production line and allows that the key point detection based veredict algorithm may be improved in future works in order to achieve a complete automation of the visual inspection process for accessories in the refrigerator.

Keywords : quality assurance; manufacturing; refrigerator; computer vision.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )