3D Transparent and Specular Object Inspection

For a long time, single-pixel imaging has been a popular research area. With the aid of this imaging modality, many methods were developed for reconstructing 3D models of opaque objects. However, three-dimensional imaging of transparent and specular objects is necessary for applications such as automatic inspection, oceanology, fluid mechanics, and computer graphics. The proposed work focuses on developing a single-pixel 3D transparent and specular object inspection system that makes use of binocular vision and polarization difference of imaging to determine the depth of the target. Within the research scope, the selection of appropriate patterns for obtaining maximum spatial information of the object is investigated. The 3D image reconstruction algorithm is also developed to deduce the depth of the object.