Industrial robot product inspection system based on 3D vision guidance

Industrial robot product inspection system based on 3D vision guidance

연구 분야: Infrastructure

학회: International Journal of Intelligent Robotics and Applications

Traditional industrial cameras adjust the focal length by rotating the zoom ring, and the process of changing the focal ring requires manual operation, but manual adjustment often brings uncertainty in time and accuracy. This paper combines the idea of 3D vision guidance to design an intelligent algorithm suitable for industrial robot product detection, so as to improve the detection rate of defective products of industrial products and improve the visual performance of industrial robots. Moreover, this paper uses image fusion to process the images obtained by multi-band illumination. Then, this paper obtains a globally clear image by image fusion, so that all the image information obtained by multi-band illumination can be completely displayed in one image, and the clarity of a single image is expanded by super-resolution, thus improving the tomographic problem. In addition, this paper carries out experimental verification by expanding the depth of field by dispersion. The experimental results show that the SSIM index of the images after super-resolution processing has been improved to a certain extent, increasing by 4.4%. Compared with the current advanced algorithm models, it has better recognition performance and system performance in product detection. Finally, this paper verifies that the industrial robot product inspection method based on 3D vision guidance is advanced, and it can be extended to other visual inspections to better serve industrial production.