Real-time implementation of a vision-aided integrated navigation system with a tiny vision processing camera module

This paper aims to implement a GPS/IMU/Vision sensor integrated navigation system which can operate in GPS-denied environments. This system is composed of a small and low-cost GPS, IMU and vision sensor. Both navigation filter and feature point tracking algorithm are loaded to Linux-based embedded processor for real-time operation. In this system, GPS/INS integrated filter and vision-based SLAM filter is used for navigation in that limited environment. With this navigation system outdoor flight test is carried out for proving its feasibility in unstructured environment.