Sensores de imagem CMOS com gradeamento metálico para detecção de distribuição angular espacial e polarização da luz

Abstract

Detecting the local angle of incidence of light is a desirable feature for common CMOS image sensors, as well as the functionality to detect the polarization of light for image reconstruction purposes in 3D and depth detection of an image. Advances in CMOS technologies in the last years have enabled integrated solutions to achieve information in the field of three-dimensional image sensors. However, it is still not viable to implement such a feature in regular CMOS image sensors due to the great number of dedicated pixels in a macropíxel to perform incident angle detection and polarization simultaneously, which implies a low resolution. In this work, a hybrid cluster with only four pixels is proposed, instead of eight pixels of previous solutions, that is able to detect local light intensity, incident angle and Stokes parameters. The technique to detect local incident angle is widely exploited in the literature. Three novelties are explored in this work, the first is the new paradigm in polarization cluster-pixel design, the second is the extended ability of metal shielded pixels to detect both the local light angle and intensity and the third is to determine the Stokes parameters through this sensor. SPICE simulation results show that the existing quadrature pixel cluster (QPC) and polarization pixel cluster (PPC) models are in accordance with experimental results presented in the literature, and thus it was possible to demonstrate similar behavior in the new hybrid polarization-quadrature pixel cluster (HPC) proposed. In addition, we work with polarized images in order to reconstruct a three-dimensional image, from the Stokes parameters.