Dense 3D depth map with DOE pattern

Author

Seung Min Choi ; Jae-chan Jeong ; Jiho Chang

Author_Institution

Electron. & Telecommun. Res. Inst., Univ. of Sci. & Technol., Daejeon, South Korea

fYear

2012

fDate

26-28 Nov. 2012

Firstpage

34

Lastpage

37

Abstract

In this paper, we try to find the answer of question, “How dense does the projected pattern have to be in order to recognize 3D fingers over 3m distance (indoor watching TV environment)”. To solve this problem, we investigate the structured light pattern projected by the laser light source and taken back through the IR camera. For 3D correspondence matching, a simple block matching algorithm from OpenCV2.4 is used. We use Canon 650D DSLR camera (removing IR-cut filter) to capture the scene, and commercial DOE product to make patterns. According to the result, at least 22.8% white dot pattern at black background is needed to recognize 3D finger shape in smart TV environment in 1296 * 864 resolution images. The experimental results could be used as such analysis or the development of new 3D sensor based structured light.

Keywords

diffractive optical elements; image matching; image resolution; infrared imaging; 3D correspondence matching; 3D fingers; Canon 650D DSLR camera; DOE pattern; IR camera; OpenCV2.4; black background; commercial DOE product; dense 3D depth map; image resolution; laser light source; projected pattern; simple block matching algorithm; smart TV environment; structured light pattern; white dot pattern; Ambient intelligence; Radio frequency; Robots; Silicon; 3D; DOE; Kinect; UI/UX; depth map; disparity map; finger; gesture recognition; smart TV; stereo vision;

fLanguage

English

Publisher

ieee

Conference_Titel

Ubiquitous Robots and Ambient Intelligence (URAI), 2012 9th International Conference on

Conference_Location

Daejeon

Print_ISBN

978-1-4673-3111-1

Electronic_ISBN

978-1-4673-3110-4

Type

conf

DOI

10.1109/URAI.2012.6462924

Filename

6462924