Article Figures & Data
Figures
- FIGURE 1
Reference frames in visual positioning; the blue stars with different shapes illustrate four feature points in the 3D space, and the round dots are their 2D projection in the image plane (the projection rays are denoted by dashed lines).
- FIGURE 2
Fundamental procedure and error space propagation of visual positioning
- FIGURE 3
An example of overexposure effect
- FIGURE 4
An example of motion blur effect; images from public data set (Seibold et al., 2017)
- FIGURE 5
An example of wrong association and outlier features; images from public data set (Geiger et al., 2012)
- FIGURE 6
Concept of the Stanford-ESA Integrity Diagram (European Space Agency, 2011)
- FIGURE 7
Different lighting conditions for local corners
- FIGURE 8
An example of initial guess impact on pose estimation; the blue circles illustrate the position of the five feature points used in the simulation.
- FIGURE 9
Potential overexposure impacts; the red circles and blue dots denote the feature point location in the image with lower and higher overexposure respectively. Image from public data set (Delmerico et al., 2019).
- FIGURE 10
Potential motion blur impacts; the blue circles denote the original corner location, while the red stars denote the corners extracted due to motion blur. Image from public data set (Delmerico et al., 2019).
- FIGURE 11
Integrity framework for visual positioning
Tables
Application Integrity risk Alert limit CAT I landing1 ≤ 10−7 ≤ 10m CAT II/III landing ≤ 10−9 ≤ 10m Autonomous driving ≤ 10−6 ≤ 2.5m Train track change ≤ 10−8 ≤ 2.5m Reference Tackled Problem Sensor Setup Integrity of Vision Error Quantification Monitored Vision Error Source Mario and Rife (2010) lane localization camera considered w/ad-hoc parameter intensity domain detection error Shytermeja et al. (2014) urban navigation camera+GNSS+IMU not considered only for GNSS vision not monitored Fu et al. (2015) landing aircraft positioning camera+GNSS partially considered addressed high-level feature domain biases Calhoun et al. (2015) relative positioning camera considered addressed position domain nominal error Zhu and Taylor (2017) correlation of error camera+IMU considered addressed position domain covariance error Yang et al. (2018) error model camera+IMU considered w/ad-hoc parameter feature association error Al Hage et al. (2019) lane localization camera+GNSS partially considered only for GNSS high-level feature domain biases Zhu et al. (2019a 2019b) error model & propagation camera considered addressed feature domain noise Zhu et al. (2020) landmark association camera considered addressed feature association error Bhamidipati and Gao (2020) urban navigation camera+GNSS considered addressed intensity domain biases Wang et al. (2020) urban navigation camera considered addressed feature domain biases
Jump to section
- Article
- Summary
- 1 INTRODUCTION
- 2 THE BASICS OF VISUAL NAVIGATION
- 3 INTRODUCTION OF INTEGRITY
- 4 RECENT DEVELOPMENTS IN VISUAL POSITIONING INTEGRITY
- 5 AN INTEGRITY DESCRIPTION FRAMEWORK FOR VISUAL POSITIONING
- 6 FUTURE TRENDS AND OPEN ISSUES
- 7 CONCLUSION
- HOW TO CITE THIS ARTICLE
- AUTHOR CONTRIBUTIONS
- CONFLICT OF INTEREST
- ACKNOWLEDGMENTS
- Footnotes
- REFERENCES
- Figures & Data
- Supplemental
- References
- Info & Metrics
Related Articles
Cited By...
- No citing articles found.