To generate high-quality video/map fusion with Enhanced Synthetic Vision System, the precise pose information should be estimated during the aircraft approach. While the accuracy of the pose data obtained by the navigation system without Ground-Based Augmentation System can’t meet the requirements for map/video registration. Fortunately, the multi-modal image sensors can be used to correct the imprecision of pose data. The traditional vision-aided pose estimation is achieved by filter method such as Kalman Filter. Due to the strong nonlinear perspective projection measurement and non-Gaussian error distribution in the estimation process, the result of the Kalman filtering method is not satisfactory. In order to solve the above problems, we propose two novel solutions, namely, Marginalized Particle Filter based and optimization-based estimation in this paper. Experimental results show that both methods have better performance on attitude estimation which is critical in the video/map fusion problem. The best final fusion is achieved through parameters calculated by optimization-based method. Because it has higher estimation accuracy compared to Marginalized Particle Filter.
Visual Aided Pose Estimation for Enhanced Synthetic Vision System
Lect. Notes Electrical Eng.
2021-10-30
11 pages
Article/Chapter (Book)
Electronic Resource
English
Visual Aided Pose Estimation for Enhanced Synthetic Vision System
British Library Conference Proceedings | 2022
|Vision-Aided Inertial Navigation for Pose Estimation of Aerial Vehicles
British Library Conference Proceedings | 2009
|Multi-Antenna Vision-and-Inertial-Aided CDGNSS for Micro Aerial Vehicle Pose Estimation
BASE | 2020
|MULTI-ANTENNA VISION-AND-INERTIAL-AIDED CDGNSS FOR MICRO AERIAL VEHICLE POSE ESTIMATION
British Library Conference Proceedings | 2020
|