光场图像深度估计是光场三维重建、目标检测、跟踪等应用中十分关键的技术.虽然光场图像的重聚焦特性为深度估计提供了非常有用的信息,但是在处理遮挡区域、边缘区域、噪声干扰等情况时,光场图像深度估计仍然存在很大的挑战.因此,提出了一种基于极平面图(Epipolar plane image,EPI)斜线像素一致性和极平面图区域差异性的深度估计算法用于解决遮挡和噪声问题.EPI斜线像素的一致性采用旋转线性算子(Spinning linear operator,SLO)的颜色熵度量,能够提高深度图边缘的准确性以及抗噪能力;EPI区域的差异性采用旋转平行四边形算子(Spinning parallelogram operator,SPO)的卡方x2度量,能够提高深度图深度渐变区域的准确性,并使用置信度加权的方法将两种度量进行融合,可以减少遮挡区域和噪声的干扰.另外,充分利用像素邻域的颜色相似性,使用引导保边滤波器和马尔科夫随机场(Markov random field,MRF)全局优化策略进行后处理,进一步减少深度图的边缘错误,得到遮挡边缘准确的深度图.在HCI光场数据集上进行了实验,并与经典光场深度估计算法进行了对比,结果表明该算法在主观质量和客观指标两方面都有明显提升. 相似文献
光场图像深度估计是光场三维重建、目标检测、跟踪等应用中十分关键的技术.虽然光场图像的重聚焦特性为深度估计提供了非常有用的信息,但是在处理遮挡区域、边缘区域、噪声干扰等情况时,光场图像深度估计仍然存在很大的挑战.因此,提出了一种基于极平面图(Epipolar plane image,EPI)斜线像素一致性和极平面图区域差异性的深度估计算法用于解决遮挡和噪声问题.EPI斜线像素的一致性采用旋转线性算子(Spinning linear operator,SLO)的颜色熵度量,能够提高深度图边缘的准确性以及抗噪能力;EPI区域的差异性采用旋转平行四边形算子(Spinning parallelogram operator,SPO)的卡方x2度量,能够提高深度图深度渐变区域的准确性,并使用置信度加权的方法将两种度量进行融合,可以减少遮挡区域和噪声的干扰.另外,充分利用像素邻域的颜色相似性,使用引导保边滤波器和马尔科夫随机场(Markov random field,MRF)全局优化策略进行后处理,进一步减少深度图的边缘错误,得到遮挡边缘准确的深度图.在HCI光场数据集上进行了实验,并与经典光场深度估计算法进行了对比,结果表明该算法在主观质量和客观指标两方面都有明显提升. 相似文献
In this paper, we design a three-dimensional micromixer based on Cantor structure. According to the principle of enhancing chaotic advection and folding fluid, it can produce better mixing performance. First we design the fractal obstacle based on the Cantor structure. We combine the obstacle with the T type microchannel. And we get multiple combinations of microchannel. We use multi-physics field simulation software COMSOL 5.2a to simulation, which is based on finite element theory. Then we analyze the mixing performance of the Imitate Cantor structure micromixer, x stands for the height of the micromixer (ICSMx) in the Re of 0.01–100 and explain the mechanisms of mixing enhancement in each structure. We compare the effect of the height of the obstacles, the effect of the spacing between the obstacles, and the effect of fractal obstacles series. By comparison, when Re is more than 50 or less than 0.1, the mixing efficiency of all micromixer can reach above 90%. Finally, we obtain a best micromixer, called Imitate Cantor structure micromixer with height 600 µm (ICSM600). The minimum mixing efficiency of ICSM600 can reach 85%, so the mixing efficiency of ICSM600 is clearly better than others.