Development of a SIFT based monocular EKF-SLAM algorithm for a small unmanned aerial vehicle

This paper describes a simultaneous localization and mapping (SLAM) algorithm using a monocular camera for a small unmanned aerial vehicle (UAV). A small U AV is attracted the attention for effective means of the collecting aerial information. However, there are few practical applications due to its small payloads for the 3D measurement. We propose extended Kalman filter (EKF) SLAM to increase UAV position and attitude data and to construct 3D terrain maps using a small monocular camera. We propose 3D measurement based on scale-invariant feature transform (SIFT) triangulation features extracted from captured images. Field-experiment results show that our proposal effectively estimates U AV position and attitude of the U AV and construct the 3D terrain map.