Real-Time MUAV Video Augmentation with Geo-information for Remote Monitoring

Micro Unmanned Aerial Vehicles (MUAV) offers a flexible and cheap way to acquire aerial images and videos, and it becomes a convenient solution to the spatial data capturing. High-resolution real-time videos broadcast by a flying MUAV can provide many applications with effective remote monitoring, but generally people can´t read other geo-information from video frames of the places where they are not familiar except the MUAV locations that are on the digital maps of the geographic information system (GIS). In order to promote the remote monitoring effect and make the data post-processing easier, this paper proposes a solution to augmenting real-time videos from the remote flying MUAV with geo-information. The solution includes four steps: 1) geo-data collection: large-scale digital maps of the MUAV monitoring area are required, and they should be converted from 2D to 3D for real-time visualization, 2) high-precision tracking for MUAV flying pose: DGPS and inertial sensors are used to capture MUAV pose in the high frequency of 20Hz, 3) video and geo-information registration: interactive technology is applied to get robust augmented results of the video, 4) interactive analysis: users query or measure places and distances directly on the video with the support of registered geo-information. We developed a prototype system including a server (temporally used as a ground station)and a terminal (i.e. MUAV platform), and the experiment was performed in the campus of Peking University, in which one could read the names of buildings and roads directly. The main advantages of the system are: 1) DGPS on the MUAV platform can provide much more stable and high-precision real-time location information than those augmented reality (AR) systems playing on the ground, as there is no interference to GPS signals from the ground objects, such as trees, buildings, and water etc., 2) Immediate geo-information can be obtained from the augmented video, 3) the traditional UAV ground stat- on is promoted by a 3D dynamic video subsystem, which can also provide query and spatial analysis.