顾键萍. 基于改进纹理表示和Mean Shift的目标跟踪[J]. 微电子学与计算机, 2013, 30(1): 161-164.
引用本文: 顾键萍. 基于改进纹理表示和Mean Shift的目标跟踪[J]. 微电子学与计算机, 2013, 30(1): 161-164.
GU Jian-ping. Object Tracking Based on Improved Texture Representation and Mean Shift[J]. Microelectronics & Computer, 2013, 30(1): 161-164.
Citation: GU Jian-ping. Object Tracking Based on Improved Texture Representation and Mean Shift[J]. Microelectronics & Computer, 2013, 30(1): 161-164.

基于改进纹理表示和Mean Shift的目标跟踪

Object Tracking Based on Improved Texture Representation and Mean Shift

  • 摘要: 本文提出了一种改进的纹理表示方法.首先对视频每帧图像进行平移不变小波变换,可以很好的提取图像的纹理和细节特征,同时能降低光照和噪声的影响,然后通过融合边缘强度来突出图像的纹理特性,把改进的纹理图像替代之前的原始帧图像.通过实验结果表明,基于改进纹理表示的算法是非常有效地,它能极大提高跟踪算法的鲁棒性.

     

    Abstract: An improved texture representation method is proposed in this essay.Translation invariant wavelet transform is performed on images from video,and then video image texture representation is obtained by the low frequency coefficients which are fused with edge intensity.The improved method can extract texture and detailed features,and it can also decrease the influences brought by illumination and noise variations.Experimental results show that the improved tracking algorithm is very effectively,it can greatly improve the robustness of the tracking algorithm.

     

/

返回文章
返回