Abstract:
A novel algorithm for recovering depth of objects from defocus images is presented, based on the anisotropic heat diffusion model. The defocus process is modeled using the model of anisotropic heat diffusion. The depth recovery problem is converted into the energy functional minimum problem with a regularization term of total variation. The depth information of the object is obtained by iterative procedures. The algorithm avoids recovering a focus image and exerting excess restrictions. Experimental results show that the algorithm is valid with small error.
Key words:
depth information,
defocus image,
anisotropic heat diffusion,
total variation
摘要:
提出一种基于各向异性热扩散方程的散焦图像深度恢复算法。利用各向异性热扩散建模散焦成像过程,将散焦图像深度恢复转化为带有整体变分正则化项的能量泛函极值问题,通过迭代获得景物的深度信息。该算法不需要恢复聚焦图像,并且未施加额外的约束条件。模拟和真实图像实验结果表明,该算法有效,且深度恢复效果优于最小二乘法。
关键词:
深度信息,
焦图像,
向异性热扩散,
体变分
CLC Number:
LIU Gong, GU Yu, CHENG Hong, HUI Sui. Novel Algorithm of Recovering Depth from Defocus Image[J]. Computer Engineering, 2010, 36(18): 169-170.
刘红, 贾郁, 程鸿, 韦穗. 一种新的散焦图像深度恢复算法[J]. 计算机工程, 2010, 36(18): 169-170.