摘要
提出了一种卫星遥感图像中具有少纹理特征的车辆弱目标检测与跟踪的交互式方法。首先,采用多边形对遥感图像中的道路区域进行人工标记,然后利用区域生长法提取道路区域。其次,融合迭代阈值法和Otsu阈值法分割道路区域图像,并通过面积约束检测车辆目标。最后,通过相邻帧图像模板匹配和距离约束实现多个车辆目标的同步跟踪。测试数据采用1米分辨率的视频卫星遥感图像,结果表明提出方法对弱小车辆目标检测的召回率为83.3%,虚警概率为2.44%。
引文
[1]朱厉洪,回征,任德锋.视频成像卫星发展现状与启示[J].卫星应用,2015,10,23-28.
[2]D.Hounam,S.Baumgartner,K.H.Bethke.An autonomous non-cooperative,wide-area Traffic Monitoring System using space-based Radar(TRAMRAD)[J].IEEE International Geoscience and Remote Sensing Symposium,20052917-2925.
[3]S.Hinz,R.Bamler and U.Stilla.Theme Issue:Airborne and Spaceborne Traffic Monitoring[J].Journal of Photo-grammetry and Remote Sensing,2006,61,135-280.
[4]张过.卫星视频处理与应用进展[J].应用科学学报,2016,34(4):361-370.
[5]G.Kopsiaftis,K.Karantzalos.Vehicle detection and traffic density monitoring from very high resolution satellite video data//[C]IEEE International Geoscience and Remote Sensing Symposium,2015,1881-1884.
[6]T.Yang,X.W.Wang,B.W Yao,et al.Small moving vehicle detection in a satellite video of an urban area[J].Sensors,2016,16,1528:1-15.
[7]袁益琴,何国金,王桂周等.背景差分与帧间差分相融合的遥感卫星视频运动车辆检测方法[J].中国科学院大学学报,2018,35(1):50-58.
[8]胡向红,陈康宁.由区域生长算法实现四边形网格划分[J].计算机辅助设计与图形学学报,2004,16(1):29-34.
[9]N.Otsu N.A threshold selection method from gray-level histogram[J].IEEE Transactions on Systems,Man,and Cybernetics,1979,9(1):62-66