The point feature, whose dynamic robustness has been widely studied, dominates in the field of visual simultaneous localization and mapping (SLAM) or visual-inertial odometry (VIO). When discussing a dynamic scene, line features are not given enough attention. This paper proposes a pre-processing step for VIO to reduce the influence of lines upon dynamic objects on system robustness and merges it into a state-of-the-art optimization-based VIO pipeline. First, it is determined whether the line feature is a potential dynamic line based upon the result of the semantic segmentation, optical flow and re-projection error. Then, instead of filtering them out, the information matrixes of these line features in the optimization function is adjusted by a weight-based method according to their tracked size. A simulated challenged visual-inertial dataset is used to evaluate the proposed algorithm against other state-of-the-art methods. The results shows that proposed method can increase robustness to dynamic scenes and make the line-based map neater and more intuitive through avoiding the drawing of dynamic line features during the mapping procedure.

The global tropospheric zenith delay grid products of VMF1 and VMF3 (Vienna mapping functions) with different resolutions are used to calculate the tropospheric zenith delay of eight IGS (International GNSS Service) stations in China, and the accuracy of the two products under different interpolation methods is analyzed. As a result, the accuracy of utilizing different interpolation methods shows no obvious differences. The interpolation accuracy of the VMF3 grid model is slightly higher than that of the VMF1, and the interpolation accuracy of tropospheric delay is related to the elevation difference of grid points. In addition, according to ERA5 (the fifth generation of the Global Climate Information Analysis data set), the atmospheric stratification tropospheric delay is obtained, and a ZTD (the zenith tropospheric delay) height change grid model is constructed using the least squares exponential fitting method. The accuracy of the model is verified using the tropospheric delay product provided by the IGS. Finally, the constructed ZTD height change grid model is used as ZTD height reduction to solve the problem of large tropospheric delay errors in the VMF interpolation when the height change is large. The model accuracy of URUM station improve from 96.47 mm.to 32.97 mm (65.82%).

The accuracy of satellite positioning results depends on the number of available satellites in the sky. In complex environments such as urban canyons, the effectiveness of satellite positioning is often compromised. To enhance the positioning accuracy of low-cost sensors, this paper combines the visual odometer data output by Xtion with the GNSS/IMU integrated positioning data output by the satellite receiver and MEMS IMU both in the mobile phone through adaptive Kalman filtering to improve positioning accuracy. Studies conducted in different experimental scenarios have found that in unobstructed environments, the RMSE of GNSS/IMU/visual fusion positioning accuracy improves by 50.4% compared to satellite positioning and by 24.4% compared to GNSS/IMU integrated positioning. In obstructed environments, the RMSE of GNSS/IMU/visual fusion positioning accuracy improves by 57.8% compared to satellite positioning and by 36.8% compared to GNSS/IMU integrated positioning.

The accuracy of satellite positioning results depends on the number of available satellites in the sky. In complex environments such as urban canyons, the effectiveness of satellite positioning is often compromised. To enhance the positioning accuracy of low-cost sensors, this paper combines the visual odometer data output by Xtion with the integrated navigation data output by the satellite receiver and MEMS IMU both in the mobile phone through adaptive Kalman filtering to improve positioning accuracy. Studies conducted in different experimental scenarios have found that in unobstructed environments, the RMSE accuracy of GNSS/IMU/visual fusion positioning improves by 50.4% compared to satellite positioning and by 24.4% compared to GNSS/IMU positioning. In obstructed environments, the RMSE accuracy of GNSS/IMU/visual fusion positioning improves by 57.8% compared to satellite positioning and by 36.8% compared to GNSS/IMU integrated positioning.