Visual 3-D SLAM from UAVs

Abstract

The aim of the paper is to present, test and discuss the implementation of Visual SLAM techniques to images taken from Unmanned Aerial Vehicles (UAVs) outdoors, in partially structured environments. Every issue of the whole process is discussed in order to obtain more accurate localization and mapping from UAVs flights. Firstly, the issues related to the visual features of objects in the scene, their distance to the UAV, and the related image acquisition system and their calibration are evaluated for improving the whole process. Other important, considered issues are related to the image processing techniques, such as interest point detection, the matching procedure and the scaling factor. The whole system has been tested using the COLIBRI mini UAV in partially structured environments. The results that have been obtained for localization, tested against the GPS information of the flights, show that Visual SLAM delivers reliable localization and mapping that makes it suitable for some outdoors applications when flying UAVs.

Audio-based localization for swarms of micro air vehicles

“…In this group of methods, localization is achieved using probabilistic techniques and by only employing on-board proprioceptive and exteroceptive sensor information. The most common examples of this approach are the vision based SLAM (Simultaneous Localization and Mapping) algorithms that mainly use an onboard camera to map features in the environment and to localize the robot [14], [15], [16]. A drawback with the local localization methods is that they mainly require a high computational power and a high data storage for operation.…”

“…This might not always be available, specially on small scale micro air vehicles. The need for real-time processing of high resolution and high frame-rate images, the dependency on illumination, visual contrast, weather conditions and the limited field of view of vision sensors, the errors caused due to the high or insufficient number of features in the images, the long displacement between loop closings and the fast dynamic nature of MAVs, are some of the major drawbacks of the visual SLAM methods for aerial robots [16].…”

Underwater Robot Swarms: Challenges and Opportunities

“…Finally, real-time SLAM has been demonstrated on-board on large outdoor UAV platforms where there is sufficient payload for high-powered computers. For instance, Artieda et al (2009) demonstrate visual 3-D SLAM on a UPM-Colibri I helicopter with a 12 kg payload. Similarly, Kim et al (2007) present results with a fusion of inertial and visual SLAM ran on-board in real-time on a Brumby MKIII UAV with 20 kg payload.…”

Vision-based absolute localization for unmanned aerial vehicles

“…While vision-based localization is well understood, it is often simultaneously associated with mapping (SLAM). However, with respect to vision-based SLAM for grounded-robot applications that has been largely explored, using such approaches for aerial vehicles and outdoor applications is still topical [7], [3], [14], [1], [5]. Furthermore, pure vision-based SLAM does not provide an absolute localization, may be subject to drift, and is prone to errors due to wrong estimation of the scale factor.…”