We propose and demonstrate an approach that uses scale invariant feature transform (SIFT) and discrete viewpoint acquisition to generate large-scale elemental image array (EIA) in integral imaging. To our knowledge, the cost of acquiring the large-scale EIA with a camera array in integral imaging system is very high. To solve this problem, we adopt the discrete viewpoint pickup platform to obtain discrete viewpoint images and a small-scale EIA. The discrete viewpoint images have enough effective pixels to form the large-scale EIA for display, but they lack accurate perspective information. In contrast, the small-scale EIA has accurate perspective information, but there are not enough effective pixels in it. Therefore, we adopt SIFT descriptors to correct the perspective information of the discrete viewpoint images according to the small-scale EIA. In this way, a large-scale EIA with accurate perspective information is obtained. The proposed method, which is not only suitable for computational integral imaging, but also for real scene acquisition and optical display, can effectively decrease costs and workload. We present the simulation and experimental results and the quality evaluation to verify the feasibility and effectiveness of the proposed generation method.

我们提出并演示了一种使用尺度不变特征变换 (SIFT) 和离散视点采集在积分成像中生成大规模元素图像阵列 (EIA) 的方法。据我们所知，在集成成像系统中使用相机阵列获得大规模环评的成本非常高。为了解决这个问题，我们采用离散视点拾取平台来获取离散视点图像和小规模的 EIA。离散视点图像有足够的有效像素形成大规模环评显示，但缺乏准确的透视信息。相比之下，小规模的EIA具有准确的透视信息，但其中的有效像素不足。因此，我们采用 SIFT 描述符根据小尺度 EIA 来校正离散视点图像的透视信息。这样就得到了具有准确透视信息的大规模环评。该方法不仅适用于计算积分成像，而且适用于实景采集和光学显示，可以有效降低成本和工作量。我们提出了模拟和实验结果以及质量评估，以验证所提出的生成方法的可行性和有效性。