Directly georeferencing terrestrial imagery using MEMS-based INS/GNSS integrated systems

Niu, Xiaoji ... [et al.]

Simultaneous developments in Global Navigation Satellite Systems (GNSS) and Inertial Navigation Systems (INS) have lead to integrated INS/GNSS systems that can provide directgeoreferencing for image-based mapping systems. Such systems directly provide the position and attitude of the camera, necessary for subsequent mapping from the imagery, rather than it indirectly being determined from object-space control points as was traditionally done. This has lead to improvements in both the efficiency and cost-effectiveness of spatial data collection. The current implementation of the navigation sensors are based on high priced, restricted handling, navigation or tactical grade IMU. These constrains limits the use of Mobile Mapping Systems (MMS) in developing countries. Low cost inertial sensors known as MEMS offer a cheap and flexible tool for direct georeferencing. In this paper, the possibility of using MEMS-based INS/GNSS systems for directly georeferencing terrestrial imagery is investigated by using backward smoothing to improve the integrated system accuracy. Field test datasets collected by land vehicles under intermediate GNSS blockage environments were processed by the proposed algorithm. Results have shown that backward smoothing can significantly improve the directgeoreferencing accuracy when compared to conventional forward Kalman filtering, especially the position accuracy during GNSS signal outages. The concept of dynamic system calibration, suitable for calibrating the MMS based on MEMS INS/GNSS integration, is presented. The achieved accuracy is promising for the direct georeferencing of some closerange portable/vehicular photogrammetry systems which require lower levels of georeference. The MEMS-based mobile mapping system has unique advantages of low-cost, small size, and no government regulation.

Event: XXIII International FIG Congress : Shaping the change

Only personal, non-commercial use of this document is allowed.

Document type:Directly georeferencing terrestrial imagery using MEMS-based INS/GNSS integrated systems (1018 kB - pdf)