• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Journal of Residuals Science & Technology >Dynamic Object Detection Based on Parallel Particle Filter and Optical flow
【24h】

Dynamic Object Detection Based on Parallel Particle Filter and Optical flow

机译：基于并行粒子滤波和光流的动态目标检测

获取原文
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Moving object detection and tracking in dynamic background has a widely applied in many fields, such as robot navigation, visual tracking for vehicle, etc. Optical flow method can be used to tracking the object in the dynamic background. While it is difficult to achieve the real-time tracking due to the traditional optical flow method involves the massive calculation and large time-consuming. Based on a distributed fusion particle filter, this paper presents a novel detection method to estimate the difference of optical flow vector between background and objective and detect the background region. In order to achieve the real time process, the image in the video is firstly divided into several parts and estimated by the particle filter respectively, and the estimated results are fused by the distributed fusion method. This method can effectively identify the background when the object the background has a significant movement. Experiment shows that the developed algorithm can detect the moving objects accurately in the dynamic background.
机译：动态背景下的运动目标检测与跟踪已在机器人导航，车辆视觉跟踪等许多领域得到了广泛的应用。光流方法可用于动态背景下的目标跟踪。传统的光流法虽然难以实现实时跟踪，但计算量大，费时。基于分布式融合粒子滤波器，提出了一种新的检测方法，用于估计背景和物镜之间的光流矢量差并检测背景区域。为了实现实时处理，首先将视频中的图像分为几部分，分别由粒子滤波器进行估计，然后通过分布式融合方法对估计结果进行融合。当物体的背景发生明显运动时，该方法可以有效地识别背景。实验表明，该算法能在动态背景下准确地检测出运动物体。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号