Indoor location method of monocular vision fusing inertial navigation information

Aiming at the problem of high positioning accuracy but unstable data sources in monocular vision positioning method, and accumulative errors in inertial measurement unit (IMU) which can obtain positioning data steadily, a monocular vision indoor positioning method is proposed, which integrates inertial navigation information. The method uses fourparameter fitting model to convert the image data into positioning and attitude data, and introduces complementary filtering to correct the gyroscope reading in the inertial data solving process. Finally, the processed data is input to extended Kalman filter. as observation value to obtain the optimal position in formation. The experimental results show that the method can effectively improve the accuracy and stability of indoor positioning.