GNSS-denied geolocalization of UAVs by visual matching of onboard camera images with orthophotos

Kinnari, Jouko; Verdoja, Francesco; Kyrki, Ville

doi:10.1109/icar53236.2021.9659333

Cited by 13 publications

(8 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To reduce the dimensionality of the problem, we assume that the roll and pitch angles of the UAV can be inferred from the direction of gravity, measurable with an IMU. Altitude is inferred as part of orthoprojection method as presented in earlier work [26]. The state is then defined as…”

Section: Methodology a Uav Localizationmentioning

confidence: 99%

“…Given a map M, an observation by the UAV, I t , and a pose hypothesis X r t , we first perform an orthoprojection of the UAV image with the method presented in earlier work [26]. The orthoprojection method in [26] performs VIO and estimates the position of tracked VIO features (landmarks) with respect to drone coordinate frame in meters, assuming sufficient excitation on inertial measurements to resolve scale [27]. By assuming the ground beneath the UAV is planar, the parameters of a plane that best fits the landmark coordinates are resolved and the UAV image is projected to a top-down view by planar homography.…”

Section: B Similarity Scoring Function Structurementioning

confidence: 99%

“…As other classical image similarity measures, we used MI, which has gained attention in other UAV localization approaches [21], [23] and is shown by Mantelli et al [6] to provide marginally superior success rate to abBRIEF. Another classical similarity measure that we used as comparison is Moravec, which we evaluated in a previous work [26].…”

Section: B Localization Under Seasonal Appearance Variationmentioning

confidence: 99%

“…In this work, the images acquired by the UAV were orthoprojected using ground-truth position information and a digital elevation model (DEM) of the environment where the flight takes place, but in a final use case, orthoprojection can be done e.g., using the method presented in earlier work [26] or by the use of calibrated downward-facing camera and an altimeter. The choice to use DEM was made to exclude the impact of possible errors in elevation estimation.…”

Section: Localization On Real Uav Datamentioning

confidence: 99%

See 3 more Smart Citations

Season-Invariant GNSS-Denied Visual Localization for UAVs

Kinnari¹,

Verdoja

Kyrki

2022

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

Localization without Global Navigation Satellite Systems (GNSS) is a critical functionality in autonomous operations of unmanned aerial vehicles (UAVs). Vision-based localization on a known map can be an effective solution, but it is burdened by two main problems: places have different appearance depending on weather and season, and the perspective discrepancy between the UAV camera image and the map make matching hard. In this work, we propose a localization solution relying on matching of UAV camera images to georeferenced orthophotos with a trained convolutional neural network model that is invariant to significant seasonal appearance difference (wintersummer) between the camera image and map. We compare the convergence speed and localization accuracy of our solution to six reference methods. The results show major improvements with respect to reference methods, especially under high seasonal variation. We finally demonstrate the ability of the method to successfully localize a real UAV, showing that the proposed method is robust to perspective changes.

show abstract

Section: Methodology a Uav Localizationmentioning

confidence: 99%

Section: B Similarity Scoring Function Structurementioning

confidence: 99%

Section: B Localization Under Seasonal Appearance Variationmentioning

confidence: 99%

Section: Localization On Real Uav Datamentioning

confidence: 99%

See 2 more Smart Citations

Season-Invariant GNSS-Denied Visual Localization for UAVs

Kinnari¹,

Verdoja

Kyrki

2022

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

show abstract

“…To solve the problem, J. Kinnari et al proposed to use the camera orthophoto method to perform image matching and measure the location information of areas without GNSS signals [9]. P. Josephy et al proposed to combine Automatic dependent surveillance broadcasting (ADS-B) multipoint positioning with satellite positioning to improve the accuracy of GNSS positioning when the satellite signal is insufficient [10].…”

Section: Introductionmentioning

confidence: 99%

Calculating the target location based on sliding fiducials

Wang

Zhu

Cao³

2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

The use of trilateration and triangulation methods to obtain location information is cumbersome and error-prone. Aiming to solve these problems, this paper designs a method of calculating target location, establishes an error model and improves the accuracy of the location calculation. It uses a laser to provide a linear relationship, sliding fiducials provide basic data, calculates the location of the target point. Experimental results show that within the measurement range of 10 meters, the average error of solving the east and west coordinates 30 times is 4.7 cm.

show abstract