Unlock instant, AI-driven research and patent intelligence for your innovation.

A P6P Camera Pose Estimation Method with Unknown Focal Length

A pose estimation and camera technology, applied in the field of computer vision, can solve problems such as inability to give accurate estimates

Active Publication Date: 2022-05-10
GUILIN UNIV OF ELECTRONIC TECH
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

EPnP and UPnP methods are linear fitting methods, and their accuracy is based on the number of key points, and when the number of key points is small, it cannot give an accurate estimate

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A P6P Camera Pose Estimation Method with Unknown Focal Length
  • A P6P Camera Pose Estimation Method with Unknown Focal Length
  • A P6P Camera Pose Estimation Method with Unknown Focal Length

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0022] The present invention will be described in detail below in conjunction with accompanying drawing and specific embodiment:

[0023] Such as figure 1 As shown, a P6P camera pose estimation method with unknown focal length of the present invention includes:

[0024] S1: Read n pairs of 2D-3D key points, select 6 pairs of them, and preprocess the information of the key points according to the model characteristics of the established rotation matrix solution model; wherein,

[0025] The method for selecting key points is:

[0026] According to the image matching result, the matching points are arranged according to the matching degree from large to small, and the first 8 points are selected as candidate points;

[0027] Take the origin of the image as the starting point, and form a vector l with each candidate point i . Select the two vectors with the largest angle as l 1 and l 2 , preferably choose the same as l 2 The angle between the vector is greater than 90 degre...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a P6P camera pose estimation method with unknown focal length, including S1: reading n pairs of 2D-3D key points, selecting 6 pairs of them, and calculating the key points according to the model characteristics of the established rotation matrix solution model S2: Make the corresponding 3D coordinate points and the first 3D coordinate point form a vector, and project it vertically onto the imaging plane; S3: Introduce the connection between the vertical projection vector and the camera projection model, and establish a The matrix equation system is used to solve the rotation matrix; S4: according to the required rotation matrix, the displacement vector is solved to obtain the relative relationship between the camera coordinate system and the world coordinate system. The present invention can estimate the pose of the camera when the image is captured.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a P6P camera pose estimation method with unknown focal length. Background technique [0002] The PnP solution method refers to the problem of solving the relative position transformation between the observed object and the camera through multiple pairs of 3D and 2D matching points under the condition of known or unknown camera internal parameters. It is a commonly used method in the front-end pose tracking part of SLAM. one. [0003] In 1981, the PnP problem was first raised. At present, the main PnP solution methods include P3P (Perspective-Three-Point, P3P for short), EPnP (Efficient PnP, EPnP for short), UPnP (Uncalibrated PnP, UPnP for short), and DLT (Direct Linear Transform, DLT for short). The P3P method refers to the complete solution to the P3P problem proposed by Gao et al. in 2003. In 2009, Vincent Lepetit et al. proposed a high-precision and fast pose estim...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/73G01C11/04
CPCG06T7/75G01C11/04G06T2207/30244
Inventor 陈紫强周秉毅谢跃雷刘庆华谢振鑫张雅琼
Owner GUILIN UNIV OF ELECTRONIC TECH