学术论文

      基于对称帧差和分块背景建模的无人机视频车辆自动检测

      Automatic vehicle detection from UAV videosbased on symmetrical frame difference and background block modeling

      摘要:
      为了从广域的视角准确全面地识别交通流信息,针对无人机视频提出了基于对称帧差和分块背景建模的车辆自动检测方法.首先,对视频图像进行4×4降维处理和灰度化处理,并人工勾勒出感兴趣区域(ROI),以降低图像维度,划定检测区域;其次,利用对称帧间差分法提取ROI中的运动目标,并在此基础上应用分块背景建模获得背景图像;然后,通过背景差分初步提取车辆信息;最后,基于形态学处理等方法消除噪声,实现车辆识别.此外,提出了针对车辆识别算法的正检率、重检率、漏检率和错检率4个评价指标.基于150帧无人机视频图像对算法进行测试,结果表明:算法具有较高的正检率(均值92.29%)、较低的漏检率(均值7.31%)与错检率(均值0.39%),而重检率为0.
      Abstract:
      In order to recognize traffic flow information correctly and comprehensively from a regional perspective, aiming at UAV (unmanned aerial vehicle) videos, an automatic vehicle detection method is proposed based on symmetrical frame difference and background block modeling.First, 4×4 dimension reduction and grayscale processing were conducted on UAV video frames, and a ROI (region of interest) was marked manually, for the purpose of reducing the image scale and specifying the detection region.Secondly, moving objects in ROI were extracted by symmetrical frame difference, and thus the background image was obtained through background block modeling.Then, vehicles were preliminarily extracted using background subtraction.Finally, noises were eliminated using techniques such as morphological processing, and vehicle recognition results were obtained.Four evaluation indices, i.e., correct detection rate, repeated detection rate, missed detection rate and false detection rate, were put forward aiming at vehicle detection algorithms.Algorithm tests were conducted on 150 frames of an UAV video.Test results show that the proposed algorithm achieves high correct detection rate (averaging 92.29%), low missed detection rate (averaging 7.31%) and false detection rate (averaging 0.39%), while the repeated detection rate is 0.
      Author: Peng Bo Cai Xiaoyu Zhang Youjie Li Shaobo
      作者单位: 重庆交通大学交通运输学院,重庆,400074
      年,卷(期): 2017, 47(4)
      分类号: U491.1
      在线出版日期: 2017年8月15日
      基金项目: 重庆市社会事业与民生保障科技创新专项资助项目,重庆市教委科学研究资助项目,重庆交通大学科研启动资助项目,重庆市科委基础科学与前沿技术研究资助项目