摘要
针对运动目标检测中的背景噪声干扰及背景抖动问题,提出了一种基于多特征融合的运动目标显著性检测方法。该方法通过对静态特征和运动特征进行融合,利用快速傅里叶变化分析其在频域上的相位谱,生成运动显著图,检测出运动目标。实验结果表明,该方法能够准确的检测出视频序列中的多个运动目标,计算简便,并且能较好的抑制背景干扰以及背景抖动的问题。
Focusing on the problem of moving targets detection such as background interference and shaking in the video sequence, a new moving target detecting method based on saliency detection is proposed. The method fuses static feature and motion feature. Analyzing the phase spectrum by Fast Fourier Transfrom to generate the saliency map. The experimental result shows that this method can accurately detect the moving targets in video sequences, and the method can be better to solve the problems like background interference, background shaking.
出处
《电子设计工程》
2016年第8期145-147,共3页
Electronic Design Engineering
关键词
特征融合
相位谱
运动显著图
快速傅里叶变换
feature fusion
phase spectrum
moving saliency map
fast Fourier transform
