Research Progress on Binocular Stereo Vision Applications

Xiaoli Yang; Yuhua Xu; Lejia Ye; Xin Zhao; Fei Wang; Zhenzhong Xiao

doi:10.3788/LOP230457

Journals >Laser & Optoelectronics Progress >Volume 60 >Issue 8 >Page 0811010 > Article

Laser & Optoelectronics Progress
Vol. 60, Issue 8, 0811010 (2023)

Research Progress on Binocular Stereo Vision Applications

Xiaoli Yang^1,†, Yuhua Xu^1,†,*, Lejia Ye¹..., Xin Zhao¹, Fei Wang² and Zhenzhong Xiao^1,**|Show fewer author(s)

Author Affiliations

¹Orbbec Inc, Shenzhen 518062, Guangdong, China

²Shenzhen Oxin Technology Co., Ltd., Shenzhen 518062, Guangdong, China

show less

DOI: 10.3788/LOP230457 Cite this Article Set citation alerts

Xiaoli Yang, Yuhua Xu, Lejia Ye, Xin Zhao, Fei Wang, Zhenzhong Xiao. Research Progress on Binocular Stereo Vision Applications[J]. Laser & Optoelectronics Progress, 2023, 60(8): 0811010 Copy Citation Text

show less

Fig. 1. Schematic diagram of FlowNetSimple and FlowNetCorr structure^[61]

Download full size

Fig. 2. Schematic diagram of GCNet structure^[70]

Download full size

Fig. 3. Schematic diagram of RAFT structure^[14]

Download full size

Dataset name	Type	Year	Resolution	Scenes	Training scenes	Testing scenes	Label type
KITTI2012^［3］	Real	2012	1240×376	389	194	195	Sparse
Middleburry^［22］	Real	2014	2948×1988	30	15	15	Dense
KITTI2015^［21］	Real	2015	1242×375	400	200	200	Sparse
Scene Flow^［56］	Synthetic	2016	960×540	39049	34801	4248	Dense
ETH3D^［133］	Real	2016	940×490	47	27	20	Dense
Falling Things^［134］	Synthetic	2018	960×540	61500			Dense
IRS^［132］	Synthetic	2019	960×540	100025	84946	15079	Dense
HR-VS^［135］	Synthetic	2019	2464×2056	780			Dense
HR-VS^［135］	Real	2019	2424×1918	33			Dense
DrivingStereo^［131］	Real	2019	1762×800	182188	174437	7751	Sparse
ApolloScape^［136］	Real	2019	3130×960	5165	4156	1009	Sparse
InStereo2K^［12］	Real	2020	1080×860	2050	2000	50	Dense
Booster^［130］	Real	2022	4112×3008	419	228	191	Dense
CRE^［16］	Synthetic	2022	1920×1080	40000			Dense

Table 1. Stereo vision dataset

Company	Camera	Depth resolution and fps	Depth accuracy	FOV /（ $°$ × $°$ ）	Depth range /m	Baseline /mm
Intel RealSense^［42］	D455	1280×720@30 848×480@90	< 2% up to 4 m	87×58	0.4-20	95
ORBBEC^［137］	Dabai Pro	640×400@30 320×200@30	$\pm$ 6 mm@1 m	67.9×45.3	0.2-2.5	40
ORBBEC^［137］	Gemini 2	1280×800@30 640×400@60 320×200@60	≤ 2%（1280×800@2m & 81% ROI）	91×66	0.15-10	50
STEREOLABS^［140］	ZED 2i	4416×1242@15 3840×1080@30 2560×720@60 1344×376@100	< 1% up to 3 m < 5% up to 15 m	110×70	0.3-20	120
DUO^［145］	DUO MLX	752×480@45 640×480@49 640×240@98 640×120@192 320×480@86 320×240@168 320×120@320		165° Wide angle lens		30
Rubedos^［138］	VIPER	up to 1280×720 up to 40		70×43	1-15	200
Carnegie Robotics^［146］	MultiSenseKS21	up to 1920×1200 Up to 30	10 m：≤1.5% 20 m：≤3%	135×84	0.5-20	210
Blaxtair^［139］	Omega	1024×512@11 512×256@18		100×70	0.5-10	100
Human+^［141］	PSP010-800	1280×800@25	10 m：≤2% 30 m：≤6%	70×50	1-30	160
Percipio.XYZ^［147］	PM801-E1	1280×960@1 640×480@1 320×240@1	3.45 mm@2500 mm	64×49	0.8-4.3	300
OAK^［142］	OAK-D-Pro	1280×800@120		80×55	0.2-35	75

Table 2. Stereo vision products

Xiaoli Yang, Yuhua Xu, Lejia Ye, Xin Zhao, Fei Wang, Zhenzhong Xiao. Research Progress on Binocular Stereo Vision Applications[J]. Laser & Optoelectronics Progress, 2023, 60(8): 0811010

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information

微信扫一扫：分享

微信扫一扫：分享