Automated Geolocation in Urban Environments Using a Simple Camera-Equipped Unmanned Aerial Vehicle: A Rapid Mapping Surveying Alternative?

Abstract

<jats:p>GNSS positioning accuracy can be degraded in areas where the surrounding object geometry and morphology interacts with the GNSS signals. Specifically, urban environments pose challenges to precise GNSS positioning because of signal interference or interruptions. Also, non-GNSS surveying methods, including total stations and laser scanners, involve time consuming practices in the field and costly equipment. The present study proposes the use of an Unmanned Aerial Vehicle (UAV) for autonomous rapid mapping that resolves the problem of localization for the drone itself by acquiring location information of characteristic points on the ground in a local coordinate system using simultaneous localization and mapping (SLAM) and vision algorithms. A common UAV equipped with a camera and at least a single known point, are enough to produce a local map of the scene and to estimate the relative coordinates of pre-defined ground points along with an additional arbitrary point cloud. The resulting point cloud is readily measurable for extracting and interpreting geometric information from the area of interest. Under two novel optimization procedures performing line and plane alignment of the UAV-camera-measured point geometries, a set of experiments determines that the localization of a visual point in distances reaching 15 m from the origin, delivered a level of accuracy under 50 cm. Thus, a simple UAV with an optical sensor and a visual marker, prove quite promising and cost-effective for rapid mapping and point localization in an unknown environment.</jats:p>

Precision mapping through an RGB-Depth camera and deep learning

“…During mapping, the RGB-D image frames are inserted to the HF-Net [5] neural network in order to extract keypoints and descriptors of the scene (feature extraction module), while the system based on keypoints and descriptors, predicts the camera pose, re-localizes the camera in case of a prediction failure and extracts new keyframes aiming to map the surroundings simultaneously (fig 2). Subsequently, multi-line convergence method localizes the markers in the scene using least squares optimization [4] while plane alignment defines the horizontal plane defining the pose of the origin marker [4]. Finally, all estimations are transferred from the initial SLAM based coordinate system, to the marker coordinate system which results in targets and point cloud coordinate estimations.…”

“…The present study, proposes a cost-effective, rapid and efficient surveying solution for GNSS-denied environments where a few minutes of walking with an RGB-D (Visual + Depth) camera on hand, are enough to map an area of interest. The proposed methodology which is based on SLAM with deep learning using the multi-line convergence (MLC) and plane alignment (PA) methods proposed in the previous stage of our work [4], is able to produce accurate coordinate estimations.…”

Real-time unmanned aerial vehicle surveying using spatial criteria: a simulated study

“…Also, Sun et al 34 developed an allin-one camera-based target detection and positioning system for fixed-wing UAVs and searchand-rescue scenarios, and they evaluated it through simulation experiments. Finally, Trigakis et al 35 demonstrated a commercial UAV that may achieve point coordinate estimation under a margin of 30 cm on a constructed point cloud. Although these research studies achieve target detection and spatial approximation in a global coordinate system, they highly depend on the used image processing approaches, the corresponding camera unit specifications and calibration procedure, and they achieve a relatively low coordinate determination accuracy.…”

“…developed an all-in-one camera-based target detection and positioning system for fixed-wing UAVs and search-and-rescue scenarios, and they evaluated it through simulation experiments. Finally, Trigakis et al 35 . demonstrated a commercial UAV that may achieve point coordinate estimation under a margin of 30 cm on a constructed point cloud.…”