Robust visual localization in changing lighting conditions

Kim, Pyojin; Coltin, Brian; Alexandrov, Oleg; Kim, H. Jin

doi:10.1109/icra.2017.7989640

Cited by 28 publications

(7 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, the proposed model which is called the semantic object conversion is designed to deal with the appearance change problem in images matching with night image [18], [26], [28]. The proposed model includes instance level segmentation method and generative adversarial network, which work as local objects detector and conversion by categories, respectively.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Development of a semantic scene conversion model for image-based localization at night

Kim

Shin

Park³

et al. 2019

Proceedings of the Conference on Research in Adaptive and Convergent Systems

View full text Add to dashboard Cite

I also cannot appreciate enough of the committee members of my thesis, Dr. Kwanghee Won, Dr. Yi Liu, and my graduate faculty representative, Dr. Lan Xu for their help on my thesis in the final defense. Due to their contributions, I was able to improve the quality of my thesis deeply and rationally. I would also like to thank the senior CCT Lab members, and I will never forget the days we studied together in the lab. Finally, I cannot express enough appreciation towards my family members for their invaluable support for me. Without their help, I was not able to continue my academic career.

show abstract

Section: Resultsmentioning

confidence: 99%

“…Image-based localization is used to figure the relative position of cameras [18], [21], [23], [26], [28]. Image-based Localization is one of the most important fields in autonomous vehicle and robotics systems since precise localization can assist vehicles to make appropriate decisions.…”

Section: Introductionmentioning

confidence: 99%

Development of a semantic scene conversion model for image-based localization at night

Kim

Shin

Park³

et al. 2019

Proceedings of the Conference on Research in Adaptive and Convergent Systems

View full text Add to dashboard Cite

show abstract

“…In the context of challenging conditions, robust VO/VSLAM remains unsolved [1,3,27]. Potential solutions to these issues can be found in [28][29][30]. These works typically specialize in a particular problem (e.g., handling high dynamic range (HDR) environment), which would introduce certain overhead under common scenarios [28].…”

Section: Monocular Visual Slam and Its Challengesmentioning

confidence: 99%

Incorporating Learnt Local and Global Embeddings into Monocular Visual SLAM

Huang¹,

Ye²,

Sun³

et al. 2021

Preprint

View full text Add to dashboard Cite

Traditional approaches for Visual Simultaneous Localization and Mapping (VSLAM) rely on low-level vision information for state estimation, such as handcrafted local features or the image gradient. While significant progress has been made through this track, under more challenging configuration for monocular VSLAM, e.g., varying illumination, the performance of state-of-the-art systems generally degrades. As a consequence, robustness and accuracy for monocular VSLAM are still widely concerned. This paper presents a monocular VSLAM system that fully exploits learnt features for better state estimation. The proposed system leverages both learnt local features and global embeddings at different modules of the system: direct camera pose

show abstract

“…The robustness of motion estimation under varying illumination is a critical issue for robotic applications. Due to the physical limitations of the generic imaging sensor, securing robustness of visual odometry methods under various lighting conditions (e.g., sunrise, sunset and night) has hardly been achieved [13]. In this respect, utilizing unconventional imaging sensors operatable under non-visible spectra has drawn attention.…”

Section: Related Workmentioning

confidence: 99%

Sparse Depth Enhanced Direct Thermal-Infrared SLAM Beyond the Visible Spectrum

Shin

Kim

2019

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

In this paper, we propose a thermal-infrared simultaneous localization and mapping (SLAM) system enhanced by sparse depth measurements from Light Detection and Ranging (LiDAR). Thermal-infrared cameras are relatively robust against fog, smoke, and dynamic lighting conditions compared to RGB cameras operating under the visible spectrum. Due to the advantages of thermal-infrared cameras, exploiting them for motion estimation and mapping is highly appealing. However, operating a thermal-infrared camera directly in existing vision-based methods is difficult because of the modality difference. This paper proposes a method to use sparse depth measurement for 6-DOF motion estimation by directly tracking under 14-bit raw measurement of the thermal camera. In addition, we perform a refinement to improve the local accuracy and include a loop closure to maintain global consistency. The experimental results demonstrate that the system is not only robust under various lighting conditions such as day and night, but also overcomes the scale problem of monocular cameras. The video is available at https://youtu.be/oO7lT3uAzLc.

show abstract

Robust visual localization in changing lighting conditions

Cited by 28 publications

References 20 publications

Development of a semantic scene conversion model for image-based localization at night

Development of a semantic scene conversion model for image-based localization at night

Incorporating Learnt Local and Global Embeddings into Monocular Visual SLAM

Sparse Depth Enhanced Direct Thermal-Infrared SLAM Beyond the Visible Spectrum

Contact Info

Product

Resources

About