论文信息 - Kinect Unleashed: Getting Control over High Resolution Depth Maps

Kinect Unleashed: Getting Control over High Resolution Depth Maps

The wide availability and economic price of Kinect popularized the concept of RGB-D cameras, which are increasingly used in a wide range of applications ranging from Human Computer Interfaces to Structural Analysis, Health Care and even Art. With over 20 million units sold, it is by far the most popular depth camera, although, being closed source, its internals are not entirely known. In this paper we focus on the PS1080 chip that executes the depth extraction algorithm inside Kinect and related devices. We propose an algorithm that can estimate the output of Kinect from the raw infrared camera data. Using this algorithm, the internal reference calibration image from Kinect can be obtained. This image allows us to model Kinect as a standard stereo camera, where alternative stereo algorithms can be used increasing the versatility of Kinect.

Rainer Stiefelhagen | Manuel Martinez | R. Stiefelhagen | Manuel Martínez

[1] Kurt Konolige,et al. Projected texture stereo , 2010, 2010 IEEE International Conference on Robotics and Automation.

[2] H. Hirschmüller. Accurate and Efficient Stereo Processing by Semi-Global Matching and Mutual Information , 2005, CVPR.

[3] Michael Riis Andersen,et al. Kinect Depth Sensor Evaluation for Computer Vision Applications , 2012 .

[4] Moonhong Baeg,et al. Spatial Uncertainty Model for Visual Features Using a Kinect™ Sensor , 2012, Sensors.

[5] Heinrich Müller,et al. Impact of Thermal and Environmental Conditions on the Kinect Sensor , 2012, WDIA.

[6] Stefan Gehrig,et al. Stereo matching in the presence of sub-pixel calibration errors , 2009, 2009 IEEE Conference on Computer Vision and Pattern Recognition.

[7] Sander Oude Elberink,et al. Accuracy and Resolution of Kinect Depth Data for Indoor Mapping Applications , 2012, Sensors.

[8] Juho Kannala,et al. Accurate and Practical Calibration of a Depth and Color Camera Pair , 2011, CAIP.

[9] K. Mimura,et al. Development of non-restrictive sensing system for sleeping person using fiber grating vision sensor , 2001, MHS2001. Proceedings of 2001 International Symposium on Micromechatronics and Human Science (Cat. No.01TH8583).