Optimal video communication strategy for intelligent video analysis in unmanned aerial vehicle applications

• 作者: Bixuan ZHANG Feng QI Jin QI Kai ZHANG Yongqiang XIE Zhongbo LI
• 来源期刊: 中国航空学报（英文版）
• 年，卷（期）: 2020,（11）
• 所属期刊栏目: Review and Article
• 页码范围: 2921-2929
• 页数: 9页
• 语种: 英文

摘要

For Unmanned Aerial Vehicles (UAV), the intelligent video analysis is a key technology in intelligent autonomous control, real-time navigation and surveillance. However, poor UAV wire-less links would degrade the quality of video communication, leading to difficulties in video anal-ysis. To meet the challenges of packet-loss and limited bandwidth in adverse UAV channel environments, this paper proposes a parameter optimization mechanism for UAV intelligent video analysis. In the proposed method, an Optimal Strategy Library (OSL) is designed to optimize the parameters for video encoding and forward error correction. Adapted to the packet-loss rate and bandwidth in practical UAV wireless network, the proposed OSL can facilitate the encoding of video sequences and the recovery of degraded videos with optimal performance. Experimental results demonstrate that the proposed solution can keep intelligent video analysis working effi-ciently with adverse UAV wireless links, and is capable of maximizing the inference accuracy of Multi-Object Tracking (MOT) algorithms in various scenarios.