基于改进帧间差分的动态目标检测算法

摘要

动态目标检测在视频监控、交通安全等领域具有广泛应用。本文提出了一种基于改进帧间差分的动态目标检测算法。首先，利用帧间差分算法提取视频帧间的差分图像，然后采用图像分割算法对差分图像进行分割，提取出目标区域。为了进一步提高检测精度，本文提出了一种改进的帧间差分算法，利用前一帧的目标区域信息对当前帧进行差分。实验结果表明，该算法在检测速度和检测精度方面均优于传统的帧间差分算法和其他常用的动态目标检测算法。

关键词：动态目标检测；帧间差分；图像分割；检测精度；检测速度

Abstract

Dynamic object detection has wide applications in video surveillance, traffic safety and other fields. In this paper, we propose a dynamic object detection algorithm based on improved frame difference. Firstly, frame difference algorithm is used to extract the difference image between video frames, and then image segmentation algorithm is used to segment the difference image to extract the target region. In order to further improve the detection accuracy, an improved frame difference algorithm is proposed, which uses the target region information of the previous frame to perform difference on the current frame. Experimental results show that the proposed algorithm is superior to traditional frame difference algorithm and other commonly used dynamic object detection algorithms in terms of detection speed and accuracy.

Keywords: dynamic object detection; frame difference; image segmentation; detection accuracy; detection speed