Position parameter estimation method based on P3P algorithm

Abstract

The present invention discloses a position parameter estimation method based on a P3P algorithm. The position parameter estimation method comprises the steps of calibrating a used camera to obtain a camera parameter; under the current used pose of the camera, imaging four characteristic points in the space and obtaining the world coordinates and the image coordinates of the four characteristic points; grouping the four characteristic points into two groups, taking the two characteristic points at the diagonal positions as the public characteristic points to form the rest two characteristic points and the public characteristic points as two sets of P3P characteristic points separately; according to the world coordinates and the image coordinates of the characteristic points in each set of P3P characteristic points, utilizing the P3P algorithm to solve a set of real solutions of the position parameters of the public characteristic points, and obtaining the two sets of real solutions corresponding to the two sets of P3P characteristic points; selecting two solutions having the largest position parameter similarity from the two sets of real solutions and verifying, and finally determining the position parameters of the public characteristic points. By determining and verifying the unique solution, the position parameter solution efficiency and accuracy are improved.

Description

technical field [0001] The invention relates to a position parameter estimation method, in particular to a position parameter estimation method based on a P3P algorithm, which belongs to the technical field of computer vision. Background technique [0002] Vision-based pose estimation technology is one of the research hotspots in the field of computer vision, and it is an important part of visual positioning and navigation technology. Vision-based pose estimation technology captures images of space cooperation targets through cameras, and solves camera pose parameters through image processing and pose estimation algorithms. It has the characteristics of moderate accuracy, low cost, independent and flexible use, and anti-electromagnetic interference. [0003] As a commonly used pose measurement method, the PnP algorithm uses n (n ≥ 3) points with known relative positions in space as feature points, collects an image by the camera, and calculates the pose information of the ca...

AI Technical Summary

Problems solved by technology

However, if there is a feature point extraction error in the image processing stage, resulting in a pose calculation error, the algorithm cannot perform self-checking

Images