RESUMO
A lightfield camera prototype is constructed by directly coupling a liquid-crystal (LC) microlens array with an arrayed photosensitive sensor for performing a LC-guided refocusing-rendering imaging attached by computing disparity map and extracting featured contours of targets. The proposed camera prototype presents a capability of efficiently selecting the imaging clarity value of the electronic targets interested. Two coefficients of the calibration coefficient k and the rendering coefficient C are defined for quantitively adjusting LC-guided refocusing-rendering operations about the images acquired. A parameter Dp is also introduced for exactly expressing the local disparity of the electronic patterns selected. A parallel computing architecture based on common GPU through the OpenCL platform is adopted for improving the real-time performance of the imaging algorithms proposed, which can effectively be used to extract the pixel-leveled disparity and the featured target contours. In the proposed lightfield imaging strategy, the focusing plane can be easily selected and/or further adjusted by loading and/or varying the signal voltage applied over the LC microlenses for realizing a rapid or even intelligent autofocusing. The research lays a solid foundation for continuously developing or upgrading current lightfield imaging approaches.