#209: Automatic video content summarization using geospatial mosaics of aerial imagery


We present a framework to summarize multiple video streams from unmanned aerial vehicles (UAVs) or drones, which have very different characteristics compared to structured commercial and consumer videos that have been widely analyzed in the past. The goal of this work is to develop scalable exploitation algorithms and tools to increase the utility, retrieval, searchability, dissemination and reuse potential of long duration surveillance videos, by providing quick dynamic geospatial mosaics and motion maps. The proposed method first generates mini-mosaics that can then be combined into geo-referenced meta-mosaics using both sensor and platform metadata associated with the video combined with low-level image analysis. The overall summarization pipeline was tested on several DARPA Video and Image Retrieval and Analysis Tool (VIRAT) datasets to evaluate effectiveness and scalability.