Modeling Kinect Sensor Noise for Improved 3D Reconstruction and Tracking

By Chuong V. Nguyen, Shahram Izadi, David R. Lovell. 2012 Second International Conference on 3D Imaging, Modeling, Processing, Visualization & Transmission, 2012

379 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/901b7fb508a98d6bc429384283567f3000bebd9b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

DepthSynth: Real-Time Realistic Synthetic Data Generation from CAD Models for 2.5D Recognition (2017, neutral): “[30] and Choo et al .”
Real-Time 3D Reconstruction in Dynamic Scenes Using Point-Based Fusion (2013, methodology): “as a function of depth for triangulation-based methods [13]).”
Reward Estimation for Variance Reduction in Deep Reinforcement Learning (2018, background): “This noise is inspired by a reward signal which is sensory based, but where sensors exhibit a Gaussian noisy distribution as in [35].”
ROBOT-CENTRIC ELEVATION MAPPING WITH UNCERTAINTY ESTIMATES (2014, methodology): “We use a sensor measurement model that was introduced by Nguyen et. al ., 9 which models the noise in depth and direction as a function of the measurement depth.”
Nonparametric Calibration for Depth Sensors (2015, positive): “[10] focus on characterizing the distortion as a function of both distance and angle of a Kinect on the observed surface and derive a noise model to filter depth maps.”
Real-time 3D reconstruction at scale using voxel hashing (2013, background): “While the quality of active infrared sensors is affected significantly in outdoor scenes, our system still manages to reconstruct large-scale outdoor scenes with fine quality.”
Voxblox: Incremental 3D Euclidean Signed Distance Fields for on-board MAV planning (2016, background): “empirically determined the RGB-D sensor model, and found that the σ of a single ray measurement varied predominantly with z(2) [19], where z is the depth of the measurement in the camera frame.”
An Efficient Sampling-Based Method for Online Informative Path Planning in Unknown Environments (2019, methodology): “Commonly used 3D sensors, such as the Microsoft Kinect or Intel Realsense, have uncertainties that scale approximately quadratically with depth z [18], [19].”
Kinect range sensing: Structured-light versus Time-of-Flight Kinect (2015, background): “In the literature there are alternative approaches using a mixed spatiotemporal analysis to derive some kind of noise statistics [42], but this approach is difficult to compare with pure temporal statistics.”