Computer Graphics
TU Braunschweig

Super Resolution for Active Light Sensor Enhancement

Super Resolution for Active Light Sensor Enhancement

This thesis provides a new approach for enhancing disparity detection of

color-depth-sensors (RGB-D-sensors). The infrared images captured by an

RGB-D-sensor are evaluated to detect subtle di erences in the light path

of the sensor's laser projection and these di erences are exported as o set


An algorithm is introduced, implemented, and tested, that is able to

detect and quantify the light path's di erences. Furthermore, an approach

to obtain a reference brightness distribution of an average spot of the RGBD-

sensor's projector is presented. The reference brightness distribution is

used to identify spot positions with sub-pixel accuracy in infrared images

captured with the RGB-D-sensor.

The system is employed to detect refractive media, being introduced

to the captured scene. It is shown that the presence of refractive media

introduced to a scene can be detected.

Author(s):Yannic Schröder
Published:March 2012
Howpublished:Bachelor thesis
Project(s): Multiple Kinect Studies  Reality CG 

  title = {Super Resolution for Active Light Sensor Enhancement},
  author = {Schr{\"o}der, Yannic},
  howpublished = {Bachelor thesis},
  month = {Mar},
  year = {2012}


  • Yannic Schröder

    Fmr. Student