Robust camera pose estimation from unknown or known line correspondences

Zhang, Xiaohu; Zhang, Zheng; You, Li; Yu, Qifeng; Ou, Jianliang

doi:10.1364/ao.51.000936

Cited by 32 publications

(16 citation statements)

References 34 publications

(66 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We can obtain n p easily from l through the known intrinsic matrix of the camera. For each pair of corresponding lines, we have one equation for t [50]: Figure 6. Geometry of camera projection model.…”

Section: Solve Translation Vectormentioning

confidence: 99%

LiDAR-Camera Calibration Using Line Correspondences

Bai

Jiang

2020

Sensors

View full text Add to dashboard Cite

In this paper, we introduce a novel approach to estimate the extrinsic parameters between a LiDAR and a camera. Our method is based on line correspondences between the LiDAR point clouds and camera images. We solve the rotation matrix with 3D–2D infinity point pairs extracted from parallel lines. Then, the translation vector can be solved based on the point-on-line constraint. Different from other target-based methods, this method can be performed simply without preparing specific calibration objects because parallel lines are commonly presented in the environment. We validate our algorithm on both simulated and real data. Error analysis shows that our method can perform well in terms of robustness and accuracy.

show abstract

Section: Solve Translation Vectormentioning

confidence: 99%

LiDAR-Camera Calibration Using Line Correspondences

Bai

Jiang

2020

Sensors

View full text Add to dashboard Cite

show abstract

“…[15] proposed more recently a PnL solution using a minimum of 3 linecorrespondences, more robust to noise, but that spans 23 solutions. [16] presented the RLPnL algorithm, requiring at least 4 line correspondences. This algorithm uses subsets of 3 line correspondences and identifies a solution in the derivatives of a 16 th order polynomial.…”

Section: Related Workmentioning

confidence: 99%

Camera Pose Estimation Based on PnL With a Known Vertical Direction

Lecrosnier

Boutteau

Vasseur

et al. 2019

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

In this paper, we address the problem of camera pose estimation using 2D and 3D line features, also known as PnL (Perspective-n-Line) with a known vertical direction. The minimal number of line correspondences required to estimate the complete camera pose is 3 (P3L) in the general case, yielding to a minimum of 8 possible solutions. Prior knowledge of the vertical direction, such as provided by common sensors (e.g. Inertial Measurement Unit, or IMU), reduces the problem to a 4 Degree of Freedom (DoF) problem and outputs a single solution. We benefit this fact to decouple the remaining rotation estimation and the translation estimation and we present a twofold contribution: (1) we present a linear formulation of the PnL problem in Plücker lines coordinates with a known vertical direction, including a Gauss-Newton-based orientation and location refinement to compensate IMU sensor noise. (2) we propose a new efficient RANdom SAmple Consensus (RANSAC) scheme for both feature pairing and outliers rejection based solely on rotation estimation from 2 line pairs. This greatly diminishes the computational cost compared to a RANSAC3 or RANSAC4 scheme. We evaluate our algorithms on synthetic data and on our own real dataset. Experimental results show state of the art results in term of accuracy and runtime, when facing 2D noise, 3D noise and vertical direction sensor noise. Index Terms-Computer Vision for Other Robotic Applications, Sensor Fusion, Localization I. INTRODUCTION Camera pose estimation consists in determining the position and the orientation of a camera with respect to a reference frame. This process requires known correspondences between real world features and their projection onto the image plane. When these features are points, we refer to the well-studied Perspective-n-Point (PnP) problem [1] [2] [3] [4] [5]. In the case of line features, we are facing the challenging, more recent and less studied Perspective-n-Line (PnL) problem. A recent review of the latter methods is presented in [6]. Once the 2D and 3D lines extraction process is completed, feature pairing is not a simple task: lines lack effective descriptors, and descriptor-based pose estimation methods are computationally more expensive.

show abstract

“…Using 3 line-correspondences, [13] present a PnL method spanning 23 solutions, but more robust to noise, and operating with a minimum of 3 lines. [14] introduce a P4L algorithm operating on 3-line subsets. They recover solutions from the derivatives of a 16 th order cost function.…”

Section: Related Workmentioning

confidence: 99%

“…[15] extend the work from [14] with ASPnL (Accurate Subset based PnL), which is outlier-sensitive, but is more accurate with a small dataset. ASPnL includes a Gauss-Newton pose refinement.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Vision based vehicle relocalization in 3D line-feature map using Perspective-n-Line with a known vertical direction

Lecrosnier¹,

Boutteau²,

Vasseur³

et al. 2019

2019 IEEE Intelligent Transportation Systems Conference (ITSC)

View full text Add to dashboard Cite

Common approaches for vehicle localization propose to match LiDAR data or 2D features from cameras to a prior 3D LiDAR map. Yet, these methods require both heavy computational power often provided by GPU, and a first rough localization estimate via GNSS to be performed online. Moreover, storing and accessing 3D dense LiDAR maps can be challenging in case of city-wide coverage. In this paper, we address the problem of camera global relocalization in a prior 3D line-feature map from a single image, in a GNSS denied context and with no prior pose estimation. We propose a dual contribution. (1) We introduce a novel pose estimation method from lines, (i.e. Perspective-n-Line or PnL), with a known vertical direction. Our method benefits a Gauss-Newton optimization scheme to compensate the sensor-induced vertical direction errors, and refine the overall pose. Our algorithm requires at least 3 lines to output a pose (P3L) and requires no reformulation to operate with a higher number of lines. (2) We propose a RANSAC (RANdom SAmple Consensus) 2D-3D line matching and outliers removal algorithm requiring solely one 2D-3D line pair to operate, i.e. RANSAC1. Our method reduces the number of iteration required to match features and can be easily modified to exhaustively test all feature combinations. We evaluate the robustness of our algorithms with a synthetic data, and on a challenging sub-sequence of the KITTI dataset. *This work is part of the COPTER project, and is co-funded by the European Union and the Région Normandie. Europe is involved in Normandy through the European Regional Development Fund (ERDF)

show abstract

Robust camera pose estimation from unknown or known line correspondences

Cited by 32 publications

References 34 publications

LiDAR-Camera Calibration Using Line Correspondences

LiDAR-Camera Calibration Using Line Correspondences

Camera Pose Estimation Based on PnL With a Known Vertical Direction

Vision based vehicle relocalization in 3D line-feature map using Perspective-n-Line with a known vertical direction

Contact Info

Product

Resources

About