User-Driven Geolocation of Untagged Desert Imagery Using Digital Elevation Models

Tzeng, Eric; Zhai, Andrew; Clements, Matthew; Townshend, Raphael J.L.; Zakhor, Avideh

doi:10.1109/cvprw.2013.42

Cited by 28 publications

(29 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Somewhat similar to Baatz et al pipeline [11], Tzeng et al [12] proposed a localization approach for desert imagery. However, instead of curvelet features, concavity-based features across query and synthetic skylines are used for matching without any use of meta-data such as GPS, FOV and focal length.…”

Section: A Related Work -Mountainous Geo-localizationmentioning

confidence: 97%

Comparison of semantic segmentation approaches for horizon/sky line detection

Ahmad

Campr

Čadík

et al. 2017

2017 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

Horizon or skyline detection plays a vital role towards mountainous visual geo-localization, however most of the recently proposed visual geo-localization approaches rely on user-in-the-loop skyline detection methods. Detecting such a segmenting boundary fully autonomously would definitely be a step forward for these localization approaches. This paper provides a quantitative comparison of four such methods for autonomous horizon/sky line detection on an extensive data set. Specifically, we provide the comparison between four recently proposed segmentation methods; one explicitly targeting the problem of horizon detection [2], second focused on visual geolocalization but relying on accurate detection of skyline [15] and other two proposed for general semantic segmentation -Fully Convolutional Networks (FCN) [21] and SegNet [22]. Each of the first two methods is trained on a common training set [11] comprised of about 200 images while models for the third and fourth method are fine tuned for sky segmentation problem through transfer learning using the same data set. Each of the method is tested on an extensive test set (about 3K images) covering various challenging geographical, weather, illumination and seasonal conditions. We report average accuracy and average absolute pixel error for each of the presented formulation.

show abstract

Section: A Related Work -Mountainous Geo-localizationmentioning

confidence: 97%

Comparison of semantic segmentation approaches for horizon/sky line detection

Ahmad

Campr

Čadík

et al. 2017

2017 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

show abstract

“…The matching process is embedded in a probabilistic framework to handle large environment. Saurer et al [180] introduce the use of a Digital Elevation Model (DEM) to perform localization in mountainous terrain [162,203,37]. Bansal and Daniilidis [15] extend this idea in urban localization to perform purely geometric VBL with images as query input and DEM of a city as database.…”

Section: Geometric Informationmentioning

confidence: 99%

A survey on Visual-Based Localization: On the benefit of heterogeneous data

Piasco

Sidibé

Demonceaux

et al. 2018

Pattern Recognition

220

112

View full text Add to dashboard Cite

We are surrounded by plenty of information about our environment. From these multiple sources, numerous data could be extracted: set of images, 3D model, coloured points cloud... When classical localization devices failed (e.g. GPS sensor in cluttered environments), aforementioned data could be used within a localization framework. This is called Visual Based Localization (VBL). Due to numerous data types that can be collected from a scene, VBL encompasses a large amount of different methods. This paper presents a survey about recent methods that localize a visual acquisition system according to a known environment. We start by categorizing VBL methods into two distinct families: indirect and direct localization systems. As the localization environment is almost always dynamic, we pay special attention to methods designed to handle appearances changes occurring in a scene. Thereafter, we highlight methods exploiting heterogeneous types of data. Finally, we conclude the paper with a discussion on promising trends that could permit to a localization system to reach high precision pose estimation within an area as large as possible.

show abstract

“…Preliminary work by Stein and Medioni (1995) focuses primarily on pre-computed panoramic skyline matching with manually extracted skylines. Tzeng et al (2013) investigate a user-aided visual localization method in the desert using DEM. Once the user marks the skyline in the query image manually, this feature is looked up in the database of panoramic skylines that is rendered from the DEM.…”

Section: Related Workmentioning

confidence: 99%

A New Method of Improving the Azimuth in Mountainous Terrain by Skyline Matching

Nagy

2020

PFG

View full text Add to dashboard Cite

Augmented reality (AR) applications have a serious problem with the accuracy of the azimuth angle provided by mobile devices. The fusion of the digital magnetic compass (DMC), accelerometer and gyroscope gives the translation and rotation of the observer in 3D space. However, the precision is not always appropriate since DMC is prone to interference when using it near metal objects or electric currents. The silhouette of ridges separates the sky from the terrain and forms the skyline or horizon line in a mountainous scenery. This salient feature can be used for orientation. With the camera of the device and a digital elevation model (DEM) the correct azimuth angle could be determined. This study proposes an effective method to adjust the azimuth by identifying the skyline from an image and matches it with the skyline of the DEM. This approach does not require manual interaction. The algorithm has also been validated in a real-world environment. ZusammenfassungEine neue Methode zur Verbesserung des Azimuts im Bergland auf Basis des Matchings der Horizontlinie. Augmented Reality (AR)-Anwendungen haben oft ein Problem mit der begrenzten Genauigkeit des Azimuts, das von mobilen Geräten geliefert wird. Die Verknüpfung eines digitalen Magnetkompasses, eines Beschleunigungsmessers und eines Vermessungskreisels liefern die Translation und Rotation eines Beobachters im 3D-Raum. Die Genauigkeit ist jedoch nicht immer brauchbar, weil ein Magnetkompass anfällig für Störungen ist, vor allem verursacht durch Metall und elektrische Ströme. Die Silhouette von Bergen bildet die Horizontlinie, die für eine Richtungsbestimmung (Orientierung) genutzt werden kann. Mit einer Kamera und einem digitalen Höhenmodell (DEM) konnte mit guter Genauigkeit das Azimut bestimmt werden. Die Untersuchung schlägt eine effektive Methode vor, um das Azimut eines Magnetkompasses durch Vergleich der Horizontlinie aus einem Bild mit der gerechneten Horizontlinie aus einem DEM zu bestimmen. Dieser Ansatz benötigt keinen manuellen Eingriff. Der Algorithmus wurde in realen Umgebungen validiert.

show abstract

User-Driven Geolocation of Untagged Desert Imagery Using Digital Elevation Models

Cited by 28 publications

References 11 publications

Comparison of semantic segmentation approaches for horizon/sky line detection

Comparison of semantic segmentation approaches for horizon/sky line detection

A survey on Visual-Based Localization: On the benefit of heterogeneous data

A New Method of Improving the Azimuth in Mountainous Terrain by Skyline Matching

Contact Info

Product

Resources

About