基于边框加权角相关的光场深度估计算法

Light field imaging can simultaneously record the intensity information and direction information of light and has the ability to estimate the scene depth. However, the accuracy of depth estimation is easily affected by light field occlusion. Therefore, this paper proposes a border-weighted angular correlation depth estimation method to solve this problem. Firstly, this method divides the light field angular domain image into four border subsets and measures the correlation of pixels in these subsets respectively to construct four cost volumes, thus solving different types of occlusions. Secondly, this method proposes a weighted fusion strategy to fuse the four cost volumes, further enhancing the robustness of the algorithm while retaining its anti-occlusion ability. Finally, the fused cost volume is optimized by guided filtering to improve the accuracy of depth estimation. The experimental results show that the proposed method is superior to the existing methods in terms of quantitative indicators. At the same time, in the absolute depth measurement experiment, the proposed method can achieve high-precision measurement.

光场成像能同时记录光线的强度信息与方向信息且具备估计场景深度的能力。然而,深度估计的精度却容易受光场遮挡的影响。因此,本文提出一种边框加权角相关的深度估计方法来解决该问题。首先,该方法将光场角度域图像分成四个边框子集并分别度量这些子集中像素的相关性来构建四个代价体积,以此解决不同类型的遮挡。其次,该方法提出加权融合策略来融合四个代价体积,进一步增强算法的鲁棒性,同时保留算法的抗遮挡能力。最后,融合后的代价体积利用引导滤波对其进行优化,以提升深度估计的精度。实验结果表明,提出的方法在量化指标上优于现有的方法。同时,在绝对深度测量实验中,提出的方法能实现高精度的测量。