Monocular urban localization using street view

Li, Yu; Joly, Cyril; Bresson, Guillaume; Moutarde, Fabien

doi:10.1109/icarcv.2016.7838744

Cited by 15 publications

(18 citation statements)

References 22 publications

(25 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This city-scale localization has an accuracy that ranges from more than 1 meter to 12 meters. In [197], visual bag-of-words methods are employed to build two dictionaries from Street View images using SIFT and MSER (Maximally Stable Extremal Regions) detectors so as to have both local and regional feature descriptors. Based on these, the closest Street View from a real image can be recovered.…”

Section: Localization In Existing Mapsmentioning

confidence: 99%

Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving

Bresson

Alsayed

et al. 2017

IEEE Trans. Intell. Veh.

Self Cite

703

366

View full text Add to dashboard Cite

Abstract-In this article, we propose a survey of the Simultaneous Localization And Mapping field when considering the recent evolution of autonomous driving. The growing interest regarding self-driving cars has given new directions to localization and mapping techniques. In this survey, we give an overview of the different branches of SLAM before going into the details of specific trends that are of interest when considered with autonomous applications in mind. We first present the limits of classical approaches for autonomous driving and discuss the criteria that are essential for this kind of application. We then review the methods where the identified challenges are tackled. We mostly focus on approaches building and reusing long-term maps in various conditions (weather, season, etc.). We also go through the emerging domain of multi-vehicle SLAM and its link with self-driving cars. We survey the different paradigms of that field (centralized and distributed) and the existing solutions. Finally, we conclude by giving an overview of the various largescale experiments that have been carried out until now and discuss the remaining challenges and future orientations.

show abstract

Section: Localization In Existing Mapsmentioning

confidence: 99%

Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving

Bresson

Alsayed

et al. 2017

IEEE Trans. Intell. Veh.

Self Cite

703

366

View full text Add to dashboard Cite

show abstract

“…Training is conducted with gradient descent using the Adam optimizer [9] with a learning rate of 10 −5 and a batch size of 80 samples during 500 epochs. We compare the results obtained by our end-to-end pose regressor with our previous approach based on handcrafted features [21]. The latter uses a bag-of-words approach followed by a feature matching between the current image and the closest panoramas found.…”

Section: Experiments and Resultsmentioning

confidence: 99%

“…• The validation of the approach with real data acquired in urban environments. • A comparison of the results with a previously developed approach that uses handcrafted features [21]. The rest of this paper is divided as follows: Section II exposes the related work regarding the use of existing information sources for visual localization as well as endto-end pose regression with CNNs.…”

Section: Introductionmentioning

confidence: 99%

Urban Localization with Street Views using a Convolutional Neural Network for End-to-End Camera Pose Regression

Bresson

Joly

et al. 2019

2019 IEEE Intelligent Vehicles Symposium (IV)

Self Cite

View full text Add to dashboard Cite

“…Görüntülerin tümüne odaklanılan bu çalışmalarda farklı açılardan benzer yerler için elde edilmiş panoramik olmayan görüntüler panoramik sokak görüntüleri ile eşleştirilmektedir. Bu çalışmalardaki panoramik görüntüler için Google Sokak Görüntüleri kullanılmış, test görüntüleri de panoramik olmayan kameralardan elde edilmiştir [17][18][19]. Majdik bir mikro insansız hava aracı üzerindeki panoramik olmayan bir kameradan elde ettiği görüntüleri önceden kaydedilmiş panoramik sokak görüntüleri ile eşleştirmiştir [17].…”

Section: Gi̇ri̇ş (Introduction)unclassified

“…Agarwal'a benzer bir yaklaşım izlese de, Yu'nun önerdiği çözümde odometri ve GPS kullanımına gerek bulunmayıp, doğrultulmuş sokak görüntüleri SIFT ve MSER görsel kelimeleri ile eğitilmiş, EPnP-RANSAC kullanılarak poz kestirimi gerçekleştirilmiştir. 300 metreye yakın bir cadde üzerinde gerçekleştirilen çalışma sonucunda ortalama 6,5 metre hata ile konum kestirimi yapılmıştır [19].…”

Section: Gi̇ri̇ş (Introduction)unclassified

Panoramik görüntülerde gökyüzü sınır çizgisini kullanarak hareketli robotlar için yer kestirimi

Koku¹

2019

Gazi Üniversitesi Mühendislik Mimarlık Fakültesi Dergisi

View full text Add to dashboard Cite

Highlights:Graphical/Tabular Abstract  Signal extraction from skylines in panoramic images to be used in localization  Use of skyline in mobile robot localization  Applicability of the proposed method on small-scale mobile robots operating outdoors Figure A. Skylines found on photos that are taken from the same spot at different times Purpose: Variety of small size and low-cost panoramic cameras are available on the market. The purpose of this work is to investigate the usability of these panoramic cameras for localization purposes especially in small-scale outdoor mobile robots. It should be noted that many mobile robots are territorial (i.e. operate in a bounded region), hence reliable map-based local navigation enhances the usability of such robots. Figure A illustrates the observation that motivates this research: Skylines on panoramic images that are captured from the same reference point at different times under different lighting conditions turn out to be very similar. Using the obvious fact that, skyline will gradually change as one moves away from this reference point, the skyline obtained at the reference point can be used as a signature signal that defines / identifies this point. Hence, skylines can be used for localization. In this study, a way of extracting signature signals from panoramic images is proposed. Signals extracted from consecutive points on a line is used to form a map. Once the map is formed, test images that are around this line are acquired at different times under different lighting conditions. The skyline signal extracted from a test image is then compared to the ones on the map using a proposed similarity measure. The map image that yields the maximum similarity that is above a certain threshold is associated with the test image and test image is considered to be localized. The study also investigates the effect of image resolution on both localization performance and computation time. Along with practical guidelines for applying the proposed method in practice, test images and source codes are provided as well. Theory and Methods: Results:Through rigorous experiments (two of which are given in the paper) it is shown that if a region is properly mapped, sub-meter localization accuracy is possible. Conclusion:What differentiates this work from others in the literature is not that skyline is used for localization purposes, but it is shown that using readily available and low cost panoramic cameras sub-meter localization can be achieved, hence, the proposed method can be used in mobile robots. Given the small size of these cameras, proposed method offers a practical localization method for small-scale outdoor robots. The dependency of the skyline being detectable by the preferred camera imposes the limitation that proposed method cannot be used during the night.Advances in robotics liberated robots from factory floors by the end of 20th century. Use of robots in our daily lives is only expected to increase in time. Robots, while relieving us of the burden of tedious, hard and dang...

show abstract

Monocular urban localization using street view

Cited by 15 publications

References 22 publications

Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving

Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving

Urban Localization with Street Views using a Convolutional Neural Network for End-to-End Camera Pose Regression

Panoramik görüntülerde gökyüzü sınır çizgisini kullanarak hareketli robotlar için yer kestirimi

Contact Info

Product

Resources

About