Incremental translation averaging method, system and device

Abstract

The invention belongs to the technical field of recovering structures from motion in three-dimensional reconstruction, and specifically relates to an incremental translational averaging method, system, and equipment, aiming to solve the problems of high complexity, low accuracy, and poor robustness of existing translational averaging methods. question. The method includes constructing an epipolar geometric graph; constructing a set of camera quadruples, selecting an initial camera quadruple as an initial seed view based on local optimization; constructing a third set of vertices; forming a camera triplet, and adopting the next step based on a weighted support set. The optimal view selection strategy determines the incremental order of vertices; performs weighted local optimization / weighted global optimization on the current estimated absolute position; performs re-translational averaging on all vertices with estimated absolute positions after global optimization. The invention reduces the complexity of the translation averaging method, and improves the accuracy and robustness of the absolute position estimation.

Description

technical field [0001] The invention belongs to the technical field of structure recovery from motion in three-dimensional reconstruction, and in particular relates to an incremental translation averaging method, system and equipment. Background technique [0002] Structure from motion is a key step in image-based large-scale scene 3D reconstruction, which has developed rapidly in recent years. Its input is image feature matching, and its output is camera absolute pose and scene structure. According to the different initialization methods of the camera pose, the recovery structure from motion can be roughly divided into two types: incremental and global. The incremental method completes the initialization of camera pose and scene structure through iterative camera pose estimation and scene structure expansion. In the iterative process, in order to cope with the inevitable feature matching external value, this method also introduces random sampling Consistency (random sample...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the above problems in the prior art, that is, in order to solve the problems of high complexity, low accuracy and poor robustness of the existing translation averaging method, the first aspect of the present invention proposes an incremental translation averaging method method, which is applied to the solution of the absolute position of the camera in the global structure from motion recovery, the method includes:

Images