Thomas Köppel
Supervisor(s): Dr. Stefan Ohrhallinger
TU Wien
Abstract: Depth maps or point clouds extracted by depth sensors are
prone to have errors. The goal of this work is the extraction
of these errors (the noise) and a statistical estimation using
a noise model. We have combined the sensor noise model
described in this work and the noise model described in
the work by Grossmann et al., generating a single noise
model allowing a prediction of the amount of noise in specific
areas of an image at a certain distance and rotation.
We have conducted two test setups and measured the noise
from 900 mm to 3.100 mm for the generation of the noise
models. The test setup of this work focuses on determining
the noise in X, Y and Z direction, covering the whole
frustum of the respective depth sensor. Z noise was measured
against a wall and X and Y noises were measured using
a 3D chequerboard that was shifted through the room,
allowing the above-mentioned coverage of the whole frustum.
Along the edges of the cells of the chequerboard, the
X and Y noise was measured. The combined model was
evaluated by using a solid cube to classify the quality of
our noise model. The estimation of the noise is important
for applications like robot navigation that use data from
depth sensors or when reconstructing a 3D scene captured
by a depth sensor.
Keywords: 3D Reconstruction, Image Processing
Full text: Year: 2018
