Estimation of the camera pose from image point correspondences through the essential matrix and convex optimization

AbstractIn the literature of computer vision, computer graphics and robotics, the use of quaternions is exclusively related to 3D rotation representation and interpolation. In this research we found how epipoles in multi-camera systems can be used to represent camera poses in the quaternion domain. The rotational quaternion is decomposed in two epipole rotational quaternions and one z axis rotational quaternion. Quadratic form of the essential matrix is also related to quaternion factors. Thus, five pose parameters are distributed into three independent rotational quaternions resulting in measurement error separation at camera pose identification and greater flexibility at virtual camera animation. The experimental results refer to the design of free viewpoint television.

Download Full-text

Accumulative Errors Optimization for Visual Odometry of ORB-SLAM2 Based on RGB-D Cameras

ISPRS International Journal of Geo-Information ◽

10.3390/ijgi8120581 ◽

2019 ◽

Vol 8 (12) ◽

pp. 581 ◽

Cited By ~ 3

Author(s):

Jiangying Qin ◽

Ming Li ◽

Xuan Liao ◽

Jiageng Zhong

Keyword(s):

Image Matching ◽

Accurate Determination ◽

Visual Odometry ◽

Camera Motion ◽

Feature Points ◽

Camera Tracking ◽

Essential Matrix ◽

Approximate Nearest Neighbors ◽

Camera Pose ◽

Simultaneous Location And Mapping

Oriented feature from the accelerated segment test (oFAST) and rotated binary robust independent elementary features (rBRIEF) SLAM2 (ORB-SLAM2) represent a recognized complete visual simultaneous location and mapping (SLAM) framework with visual odometry as one of its core components. Given the accumulated error problem with RGB-Depth ORB-SLAM2 visual odometry, which causes a loss of camera tracking and trajectory drift, we created and implemented an improved visual odometry method to optimize the cumulative error. First, this paper proposes an adaptive threshold oFAST algorithm to extract feature points from images and rBRIEF is used to describe the feature points. Then, the fast library for approximate nearest neighbors strategy is used for image rough matching, the results of which are optimized by progressive sample consensus. The image matching precision is further improved by using an epipolar line constraint based on the essential matrix. Finally, the efficient Perspective-n-Point method is used to estimate the camera pose and a least-squares optimization problem is constructed to adjust the estimated value to obtain the final camera pose. The experimental results show that the proposed method has better robustness, higher image matching accuracy and more accurate determination of the camera motion trajectory.

Download Full-text

Learning Two-View Correspondences and Geometry via Local Neighborhood Correlation

Entropy ◽

10.3390/e23081024 ◽

2021 ◽

Vol 23 (8) ◽

pp. 1024

Author(s):

Luanyuan Dai ◽

Xin Liu ◽

Jingtao Wang ◽

Changcai Yang ◽

Riqing Chen

Keyword(s):

Pose Estimation ◽

Contextual Information ◽

Local Region ◽

K Nearest Neighbor ◽

Outlier Rejection ◽

Essential Matrix ◽

Camera Pose Estimation ◽

Local Neighborhood ◽

Camera Pose ◽

Neighborhood Correlation

Seeking quality feature correspondences (also known as matches) is a foundational step in computer vision. In our work, a novel and effective network with a stable local constraint, named the Local Neighborhood Correlation Network (LNCNet), is proposed to capture abundant contextual information of each correspondence in the local region, followed by calculating the essential matrix and camera pose estimation. Firstly, the k-Nearest Neighbor (KNN) algorithm is used to divide the local neighborhood roughly. Then, we calculate the local neighborhood correlation matrix (LNC) between the selected correspondence and other correspondences in the local region, which is used to filter outliers to obtain more accurate local neighborhood information. We cluster the filtered information into feature vectors containing richer neighborhood contextual information so that they can be used to more accurately determine the probability of correspondences as inliers. Extensive experiments have demonstrated that our proposed LNCNet performs better than some state-of-the-art networks to accomplish outlier rejection and camera pose estimation tasks in complex outdoor and indoor scenes.

Download Full-text

Analysis of STEM micrographs of thick objects

Proceedings, annual meeting, Electron Microscopy Society of America ◽

10.1017/s0424820100081395 ◽

1978 ◽

Vol 36 (2) ◽

pp. 106-107

Author(s):

S. Golladay

Keyword(s):

Inelastic Scattering ◽

Multiple Scattering ◽

Mass Distribution ◽

Cross Sections ◽

Phase Contrast ◽

Image Point ◽

Contrast Effects ◽

Total Cross Sections ◽

Elastic And Inelastic Scattering

The theory of multiple scattering has been worked out by Groves and comparisons have been made between predicted and observed signals for thick specimens observed in a STEM under conditions where phase contrast effects are unimportant. Independent measurements of the collection efficiencies of the two STEM detectors, calculations of the ratio σe/σi = R, where σe, σi are the total cross sections for elastic and inelastic scattering respectively, and a model of the unknown mass distribution are needed for these comparisons. In this paper an extension of this work will be described which allows the determination of the required efficiencies, R, and the unknown mass distribution from the data without additional measurements or models. Essential to the analysis is the fact that in a STEM two or more signal measurements can be made simultaneously at each image point.

Download Full-text

New polarized-light microscope for fast and orientation independent measurement of birefringent fine structure

Proceedings, annual meeting, Electron Microscopy Society of America ◽

10.1017/s0424820100146254 ◽

1993 ◽

Vol 51 ◽

pp. 82-83

Author(s):

Rudolf Oldenbourg

Keyword(s):

Light Microscope ◽

Polarized Light ◽

Video Camera ◽

Limited Range ◽

Image Point ◽

Computer Assisted ◽

Optical Modulators ◽

Anisotropic Structures ◽

Long Time ◽

Computer Assisted Image

The polarized light microscope has the unique potential to measure submicroscopic molecular arrangements dynamically and non-destructively in living cells and other specimens. With the traditional pol-scope, however, single images display only those anisotropic structures that have a limited range of orientations with respect to the polarization axes of the microscope. Furthermore, rapid measurements are restricted to a single image point or single area that exhibits uniform birefringence or other form of optical anisotropy, while measurements comparing several image points take an inordinately long time.We are developing a new kind of polarized light microscope which combines speed and high resolution in its measurement of the specimen anisotropy, irrespective of its orientation. The design of the new pol-scope is based on the traditional polarized light microscope with two essential modifications: circular polarizers replace linear polarizers and two electro-optical modulators replace the traditional compensator. A video camera and computer assisted image analysis provide measurements of specimen anisotropy in rapid succession for all points of the image comprising the field of view.

Download Full-text