Based On Inertial Measurement Unit And Visual Measurement Fusion Algorithm In 3D Space

Virtual reality(VR)is one of the most important technologies.IMU is a core sensor component of virtual reality technology.The integration of IMU and machine vision is the main research content of this paper.In real applications,due to its own characteristics caused by the offset error,which will accumulate over time,IMU eventually generates distortion.This problem has been plagued by many domestic and foreign VR equipment manufacturers.In this context,the relevant concepts related to IMU correction comes into being,and its goal was to use the external information as a reference to update the state of the IMU and to reduce and eliminate the cumulative error of the IMU.At present,IMU correction algorithm has been widely used in aerospace,navigation equipment,virtual reality and other fields.This paper first discusses the relevant theoretical and technical basis of IMU inertial measurement and visual measurement,as well as domestic and foreign related research results in this field.Then we respectively implement IMU inertia measurement and visual measurement.Finally,the results of the two measurements are fused to obtain the corrected IMU inertial measurement results.The main work of this paper is divided into the following three aspects:(1)IMU inertial combination algorithm based on gyroscope,accelerometer and magnetometer: The basic IMU inertial measurement unit is obtained by combining three kinds of inertial components of gyroscope,accelerometer and magnetometer.This kind of basic IMU can monitor the offset angle generated by the component when it is moving.In order to get the data which can match the visual measurement result,the shift vector is solved by Rodrigues.(2)Based on the method of visual distance measurement,the direction vector of the measured object is obtained.Firstly,the binocular camera is visually calibrated to obtain the relevant parameters such as the focal length and rotation matrix of the camera.Then,the triangulation distance theorem is used to solve the depth information,and to solve the three-dimensional coordinates of two measured markers.Finally,the direction vector between the two markers is to be synthesized.(3)IMU and visual measurement vector fusion based on Kalman filter: IMU is characterized by high output frequency and cumulative error;visual measurement is characterized by low frequency and high accuracy.The high-frequency IMU measurement is used as the state vector of the filtering system and the low-frequency visual measurement as the update vector.The final output is the corrected shift vector which is not only of high frequency,but also of low adaptive accumulative error.