Improved Reconstruction techniques with the Aid of Ray Tracing for Refractive Index Matching Free Optical Projection Tomography
Abstract
The global jewelry market demands high-quality, clear, and shiny diamonds, making their quality assessment essential. Clarity and colour are critical aspects of diamond quality evaluation and are investigated in this study.
Featuring the use of visible light sources and the detection of light transmission that mimics human vision, this research in optical projection tomography (OPT) was initially motivated by the need to complement the existing optical coherence tomography (OCT) technique. OCT is a 3D imaging method for mapping diamond inclusions based on light interferometry and the detection of back-scattered infrared light. OPT maps the internal features of transparent objects by processing projection images of a target object captured from various viewing angles. However, it suffers from refraction effects due to the use of visible light. These effects are usually mitigated by immersing or embedding the target object in a material with a refractive index closely matched to that of the object. Nevertheless, the index-matching materials used for diamonds are often toxic, complicating operational procedures and limiting the wider application of OPT within the industry. Therefore, this study aimed to use OPT for inclusion mapping of rough diamonds without employing index matching materials, correcting refraction effects computationally instead. Moreover, the use of visible light allows us to imitate human colour perception, facilitating the exploration of colour detection in rough diamonds.
Two OPT configurations were developed and examined: laser-scanning OPT, which uses a scanning laser system to capture a complete projection image point-by-point with a photodiode, a single sensing element; and telecentric OPT, which employs a colour camera with a telecentric lens for image acquisition. The latter configuration features a dual-rotation mechanism to improve ray coverage and was chosen for further study primarily due to its faster image acquisition capabilities. The in-software refraction correction is based on a ray tracing process. The light propagation at each viewing orientation is simulated from its generation, its travelling through a high-resolution object surface model created by X-ray computed tomography (XCT), to its detection. In this process, refracted ray paths are computed, and simulated projections are generated, which can be incorporated into an iterative reconstruction algorithm to reconstruct a 3D map of the object's internal features. Additionally, the acquired projection images can be further processed and analysed for diamond colour evaluation.