Reliable Fusion of Stereo Matching and Depth Sensor for High Quality Dense Depth Maps

doi:10.3390/s150820894

CORC > 计算技术研究所 > 中国科学院计算技术研究所 > 中国科学院计算技术研究所期刊论文 > 英文

	Reliable Fusion of Stereo Matching and Depth Sensor for High Quality Dense Depth Maps
	Liu, Jing 1,2; Li, Chunpeng 2; Fan, Xuefeng 1,2; Wang, Zhaoqi 2
刊名	SENSORS
	2015-08-01
卷号	15 期号:8 页码:20894-20924
关键词	stereo matching depth sensor multiscale pseudo-two-layer model segmentation texture constraint fusion move
ISSN号	1424-8220
DOI	10.3390/s150820894
英文摘要	Depth estimation is a classical problem in computer vision, which typically relies on either a depth sensor or stereo matching alone. The depth sensor provides real-time estimates in repetitive and textureless regions where stereo matching is not effective. However, stereo matching can obtain more accurate results in rich texture regions and object boundaries where the depth sensor often fails. We fuse stereo matching and the depth sensor using their complementary characteristics to improve the depth estimation. Here, texture information is incorporated as a constraint to restrict the pixel's scope of potential disparities and to reduce noise in repetitive and textureless regions. Furthermore, a novel pseudo-two-layer model is used to represent the relationship between disparities in different pixels and segments. It is more robust to luminance variation by treating information obtained from a depth sensor as prior knowledge. Segmentation is viewed as a soft constraint to reduce ambiguities caused by under- or over-segmentation. Compared to the average error rate 3.27% of the previous state-of-the-art methods, our method provides an average error rate of 2.61% on the Middlebury datasets, which shows that our method performs almost 20% better than other fused algorithms in the aspect of precision.
资助项目	National Natural Science Foundation of China[61300131] ; National Key Technology Research and Development Program of China[2013BAK03B07] ; National High Technology Research and Development Program of China (863 Program)[2013AA013902]
WOS研究方向	Chemistry ; Electrochemistry ; Instruments & Instrumentation
语种	英语
出版者	MDPI AG
WOS记录号	WOS:000360906500157
内容类型	期刊论文
源URL	[http://119.78.100.204/handle/2XEOYT63/9331]
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Liu, Jing
作者单位	1.Univ Chinese Acad Sci, Beijing 100049, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, Beijing Key Lab Mobile Comp & Pervas Device, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Liu, Jing,Li, Chunpeng,Fan, Xuefeng,et al. Reliable Fusion of Stereo Matching and Depth Sensor for High Quality Dense Depth Maps[J]. SENSORS,2015,15(8):20894-20924.
APA	Liu, Jing,Li, Chunpeng,Fan, Xuefeng,&Wang, Zhaoqi.(2015).Reliable Fusion of Stereo Matching and Depth Sensor for High Quality Dense Depth Maps.SENSORS,15(8),20894-20924.
MLA	Liu, Jing,et al."Reliable Fusion of Stereo Matching and Depth Sensor for High Quality Dense Depth Maps".SENSORS 15.8(2015):20894-20924.