Camera-inertial sensor calibration method based on partial sensor information

Abstract

The invention discloses a camera-inertial sensor calibration method based on partial sensor information. The method includes the steps that first, a calibration equation based on partial inertial sensor information is established, afterwards, the missing information of an inertial sensor is recovered first, then the relative coordinate transformation between a camera and the inertial sensor is solved, so that the calibration equation based on partial inertial sensor information is solved, and the calibration result is optimized. The camera-inertial sensor calibration problem is solved under the condition that the output information of the inertial sensor is incomplete, in the practical calibration environment, under the condition that observation errors exist, the final calibration result can be optimized through a designed optimization algorithm, and good calibration precision is achieved.

Description

technical field [0001] The invention belongs to the field of camera-inertial sensor calibration in computer vision, is applied in simultaneous localization and mapping (SLAM) technology, and specifically relates to a camera and an inertial sensor under the condition that the output information of the inertial sensor is incomplete The calibration method of coordinate system transformation between. Background technique [0002] Synchronous positioning and map creation refers to the fact that a mobile object perceives the environment through its own sensors without knowing its own position, creates a map consistent with the environment in an unknown environment, and simultaneously determines its position and posture in the map . In recent years, with the further development of technologies such as computer vision, synchronous positioning and map creation technology based on monocular vision has gradually become a research hotspot. Considering the shortcomings of a single perc...

AI Technical Summary

Problems solved by technology

At present, the low-cost and low-end inertial sensors on the market can output the roll angle (Roll angle) and the pitch angle (Pitch angle) of the three-axis direction angle with high accuracy, but the error of the output yaw angle (Yaw angle) relatively bigger

If the yaw angle information with large error output by the inertial sensor is directly used to assist the monocular vision SLAM system, not only cannot effectively reduce the cumulative error of the system, but will further increase the system error due to the introduction of a large yaw angle error

Images