Accuracy and Resolution of Kinect Depth Data for Indoor Mapping Applications
By Kourosh Khoshelham, Sander Oude Elberink. Italian National Conference on Sensors, 2012
1,800 citations recorded by the source corpus.
Consumer-grade range cameras such as the Kinect sensor have the potential to be used in mapping applications where accuracy requirements are less strict. To realize this potential insight into the geometric quality of the data acquired by the sensor is essential. In this paper we discuss the calibration of the Kinect sensor, and provide an analysis of the accuracy and resolution of its depth data. Based on a mathematical model of depth measurement from disparity a theoretical error analysis is presented, which provides an insight into the factors influencing the accuracy of the data. Experimental results show that the random error of depth measurement increases with increasing distance to the sensor, and ranges from a few millimeters up to about 4 cm at the maximum range of the sensor. The quality of the data is also found to be influenced by the low resolution of the depth measurements.
Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Geology
View Paper: https://www.semanticscholar.org/paper/63e1dffc19c3b4e99ae22ec60d10eaaafd608bcb
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
- Vladlen Koltun
- Qian-Yi Zhou
- Daniel Cremers
- Andrew J. Davison
- Thomas Whelan
- Jürgen Sturm
- Michael Kaess
- Wolfram Burgard
In Computer Vision and Pattern Recognition
- Vladlen Koltun
- Daniel Cremers
- Andrew J. Davison
- Luc Van Gool
- Qian-Yi Zhou
- Dacheng Tao
- Jitendra Malik
- Trevor Darrell
In Geology
- Vladlen Koltun
- Qian-Yi Zhou
- Daniel Cremers
- Jürgen Sturm
- Andrew J. Davison
- Michael Kaess
- Michael Firman
- Thomas Whelan
Citations in the corpus
- Evolutionary joint selection to improve human action recognition with RGB-D devices (2014, background): “Among several difficulties, lack of precision and occlusions caused by body parts or other objects present in the scene stand out (Khoshelham & Elberink, 2012; Shotton et al., 2011).”
- KinectFaceDB: A Kinect Database for Face Recognition (2014, background): “Thanks to its efficiency, low-cost, ease of RGB-D mapping, and multimodal sensing, the Kinect sensor has received vast amount of attention from diverse research communities [2], including but not limited in computer vision [3], computer graphics [4], [5], augmented reality (AR) [4], human–computer-i”
- T-LESS: An RGB-D Dataset for 6D Pose Estimation of Texture-Less Objects (2017, background): “The average absolute difference μ|δ| is less than 5mm for Carmine and 9mm for Kinect, which is near the accuracy of the sensors [28] and is relatively small compared to the size of objects.”
- 50 years of biometric research: Accomplishments, challenges, and opportunities (2016, neutral): “Consequently , most of the research in biometric recognition has rightly focused on the following two fundamental problems:”
- Communication-Efficient and Distributed Learning Over Wireless Networks: Principles and Applications (2020, methodology): “The proposed SL framework is validated by simulation with data measured in a real experiment using 60 GHz mmWave signals and two Kinect RGB-D cameras [142].”
- Robust reconstruction of indoor scenes (2015, background): “(εwas set based on typical sensor noise magnitudes [36].”
- Evaluating and Improving the Depth Accuracy of Kinect for Windows v2 (2015, background): “into the factors affecting the accuracy of the data [11], [12].”
- Elastic Fragments for Dense Scene Reconstruction (2013, background): “Obtaining a detailed three-dimensional model of an object or an environment from range images is difficult in part due to the high-frequency noise and quantization artifacts in the data [11, 27].”
- Analysis of the Accuracy and Robustness of the Leap Motion Controller (2013, methodology): “An analysis of the Kinect controller shows a standard deviation in depth accuracy of approximately 1 : 5 cm [1].”
- Dual Attention Guided Gaze Target Detection in the Wild (2021, background): “Early works [19, 21, 23] make an effort to exploit some statistically meaningful monocular cues (e.”
- Kinect range sensing: Structured-light versus Time-of-Flight Kinect (2015, methodology): “Compared to Khoshelham and Elberink [26] we observe an additional fluctuation which can be explained by the decreasing depth resolution of the Kinect which leads to a significant depth quantization for increasing distances.”