Method and device for positioning based on vision and inertia mileage

Abstract

The invention discloses a method and a device for positioning based on vision and inertia mileage. According to image feature points matching results and inertia measurement data, determining estimated pose changes of adjacent frames; when the pre-estimated pose change is smaller than a preset value, determining a to-be-solved state vector of the current frame, determining a change matrix of the to-be-solved state vector, a to-be-solved coefficient matrix and a matrix relational expression of a to-be-solved residual matrix based on the target function, solving a variation of an environment point vector and a variation of a pose vector by ignoring a high-order term of the matrix relational expression, and finally determining the pose of the current frame. Under the condition that the estimated pose change is small, that the pose change is linear and can be obtained through derivation according to a coefficient matrix, a state vector and a residual error of the previous frame is determined, and therefore an incremental expression of the coefficient matrix and the residual error is constructed, and all the state vectors are obtained by solving an incremental equation; the coefficientmatrix and residual error of the current frame do not need to be calculated, calculation complexity is reduced, and positioning efficiency is improved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a method and device for positioning based on vision and inertial mileage. Background technique [0002] Passive positioning technology refers to the method that the terminal determines its own position according to the data collected by itself without relying on other positioning devices. Wherein, other positioning devices generally refer to positioning satellites, for example, satellites of the global positioning system. At present, a main solution in the passive positioning technology is positioning based on visual and inertial odometry (Visual-Inertial Odometry, VIO), which is generally called a visual-inertial odometry positioning method. [0003] In the prior art, the visual-inertial odometry positioning method is usually implemented based on a system including front-end and back-end equipment. Among them, the front end is a device that needs to be positioned. ...

AI Technical Summary

Problems solved by technology

[0005] However, since the state vector usually contains information such as image feature information, IMU data information, and environmental features, it is very complicated to calculate the state vector based on the coefficient matrix and residuals, which requires a relatively powerful computing power for the server to meet the demand.

However, if the frames involved in the calculation are reduced, the positioning accuracy will decrease.

Images