基于多特征融合的运动阴影去除算法
|
摘要
在运动目标检测中,由于运动阴影与前景物体具有相似的运动特征,传统的方法很难区分运动对象及其阴影。为了解决这一问题,提出了一种基于多特征融合的阴影去除算法。在HSV颜色空间下利用阴影所具有的色度不变性特征对前景图像进行预处理并得到阴影的候选区域;利用小波变换分析候选区域和与之对应背景区域纹理特征的相似程度,去除其中的相似成分得到最终的检测结果。实验结果表明,提出的方法在不同场景下能够有效地去除运动阴影区域。
In the field of motion detection,the traditional method is difficult to distinguish the moving object and its shadow due to the similar motion features between the moving shadow and the foreground object.In order to solve this problem,a novel shadow removal algorithm based on multi-feature fusion is proposed in this paper.Firstly,in the HSV color space,we use chromaticity invariance feature of shadow to obtain the candidate region of shadow.Then,we use the wavelet transform to analyze the texture similarity between the candidate region and background region,and remove the similar wavelet components to obtain the final detection result.The experimental results show that our method can effectively remove moving shadow in different scenes.
In the field of motion detection,the traditional method is difficult to distinguish the moving object and its shadow due to the similar motion features between the moving shadow and the foreground object.In order to solve this problem,a novel shadow removal algorithm based on multi-feature fusion is proposed in this paper.Firstly,in the HSV color space,we use chromaticity invariance feature of shadow to obtain the candidate region of shadow.Then,we use the wavelet transform to analyze the texture similarity between the candidate region and background region,and remove the similar wavelet components to obtain the final detection result.The experimental results show that our method can effectively remove moving shadow in different scenes.
引文
[1]Cucchiara R,Grana C,Piccardi M,et al.Improving shadow suppression in moving object detection with HSV color information[C]∥Proceedings of 2001IEEE Intelligent Transportation Systems.IEEE,2001:334-339.
[2]Sun B,Li S.Moving cast shadow detection of vehicle using combined color models[C]∥2010Chinese Conference on Pattern Recognition(CCPR).IEEE,2010:1-5.
[3]Leone A,Distante C,Buccolieri F.Shadow Detection for moving objects based on texture analysis[J]∥Pattern Recognition,IEEE,2007:1222-1233.
[4]曹健,陈红倩,张凯.结合区域颜色与纹理的运动阴影检测方法[J].机器人,2011,33(5):628-633.
[5]Mallat S.A Theory for Multi-resolution Signal Decomposition:The Wavelet Representation[C]∥Transctions on Partern Analysis and Machine Intelligence.1989:674-693.
[6]William R Zetrier,John Huffman and David C P Linden.Application of Compactly Supported Wavelets to Image Compression[C]∥SPIE Image Processing Algorithms and Techniques.1990:150-161.
[7]Olivier Barnich and M Van Droogenbroeck.ViBe:A universal background subtraction algorithm for video sequences[C]∥Transactions on Image Processing.IEEE,2011,20(6):1709-1724.
[2]Sun B,Li S.Moving cast shadow detection of vehicle using combined color models[C]∥2010Chinese Conference on Pattern Recognition(CCPR).IEEE,2010:1-5.
[3]Leone A,Distante C,Buccolieri F.Shadow Detection for moving objects based on texture analysis[J]∥Pattern Recognition,IEEE,2007:1222-1233.
[4]曹健,陈红倩,张凯.结合区域颜色与纹理的运动阴影检测方法[J].机器人,2011,33(5):628-633.
[5]Mallat S.A Theory for Multi-resolution Signal Decomposition:The Wavelet Representation[C]∥Transctions on Partern Analysis and Machine Intelligence.1989:674-693.
[6]William R Zetrier,John Huffman and David C P Linden.Application of Compactly Supported Wavelets to Image Compression[C]∥SPIE Image Processing Algorithms and Techniques.1990:150-161.
[7]Olivier Barnich and M Van Droogenbroeck.ViBe:A universal background subtraction algorithm for video sequences[C]∥Transactions on Image Processing.IEEE,2011,20(6):1709-1724.