Features Detection-Based Computer Vision of Transparent 3D Inkjet Printed Parts

This paper reports on the development of a detection and recognition-based computer vision algorithm of features in transparent polymer parts. In particular , transparent polymer parts fabricated by a 3D inkjet printing process with different features from a few millimeters down to a sub-millimeter range are used for validation of the proposed algorithm. Optical images were captured using a digital microscope, by which special attention was paid to the influence factor of light. These images pass through a detection algorithm to identify and analyze the features. Detection results were validated to determine the accuracy. It has been found that the accuracy of the detection algorithm is adequate to detect errors caused by the inkjet printing process with an accuracy in the range of few hun-dredths of a millimeter. The results show that the maximum error caused by the detection is in the range of a tenth of a millimeter. However, lighting and transparency of the printed part, the quality of the taken image, and the fusion of some printed features together with which blurred edges can be detected are the main challenges for an accurate detection of the features in transparent printed parts. ... mehrFinally, it is worth stating that there still a wide margin of improvement to obtain, and further development is still required.