基于导向滤波的深度图优化方法  被引量:2

Optimization Method of Depth Image Based on Guided Filtering

在线阅读下载全文

作  者:万众[1] 熊智[1] 许建新[1] 李一博[1] WAN Zhong;XIONG Zhi;XU Jian-xin;LI Yi-bo(Nanjing University of Aeronautics and Astronautics, Nanjing 211106)

机构地区:[1]南京航空航天大学,南京211106

出  处:《导航与控制》2018年第1期49-55,共7页Navigation and Control

基  金:国家自然科学基金(编号:61673208;61533008;61533009;61374115);航空科学基金(编号:20165552043);中央高校基本科研业务费专项资金资助(编号:NZ2016104;NS2017016)

摘  要:很多低成本设备输出的深度图存在明显的边缘不匹配、深度信息缺失导致孔洞等问题,而现有的优化算法实时性差,提出的基于导向滤波的深度图优化方法可以兼顾实时性和视觉效果。首先,采用基于单尺度的Retinex方法对配准的灰度图像进行增强处理,消除光照阴影等导致的虚假边缘,增强真实边缘。然后,将处理后的灰度图像作为引导基础,通过具有边缘保持能力的导向滤波器优化深度图像,实现边缘保持的同时填充孔洞。最后,通过标准数据库和实际深度图进行实验验证。结果表明,处理后的深度图能够很好地反映基本形态,兼具实时运算竞争力。Edge mismatching and holes which missing depth leads to often occur in the depth image obtained by lowcost equipments,and the existing optimization algorithms have poor real-time performance. A real-time optimization algorithm based on guided filtering is proposed in this paper,which has great performance both in time cost and visual effects.First,this paper uses the single scale Retinex method to enhance the registered gray image,which can eliminate false edges caused by shadow and enhance the real edge. Then,taking the gray image as a guide image,optimize the depth image using guided filter with edge preserving ability to maintain the edges and fill holes. Finally,this paper tests the proposed method on both the standard database and the actual depth map. Experiment results show that the processed depth map can well maintain the real shape with real-time competitiveness on computing.

关 键 词:深度图优化 导向滤波 图像增强 边缘保持 

分 类 号:TN317.4[电子电信—物理电子学]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象