抑制船尾拖纹的船舶显著性视频检测算法
详细信息 查看官网全文
- 英文论文题名：Stern Ripples-Driven Ship Detection Algorithm in Inland Waterway Based on Visual Saliency
- 论文作者：李昀峰 ; 周东 ; 阮雅端 ; 陈林凯 ; 陈启美
- 英文论文作者：LI Yun-feng ; ZHOU Dong ; RUAN Ya-duan ; CHEN Lin-kai ; CHEN Qi-mei ; School of Electronic Science and Engineering ; Nanjing University ; College of Computer Engineering ; Jiangsu University of Technology
- 年：2016
- 作者机构：南京大学电子科学与工程学院;江苏理工学院计算机工程学院;
- 论文关键词：船尾拖纹 ; 显著性 ; 超像素分割 ; GrabCut算法
- 英文论文关键词：stern ripples ; saliency ; super-pixel segmentation ; GrabCut algorithm
- 会议召开时间：2016-06-24
- 会议录名称：2016年全国通信软件学术会议程序册与交流文集
- 语种：中文
- 分类号：TP391.41
- 学会代码：ZGTH
- 会议名称：2016年全国通信软件学术会议
- 会议地点：中国陕西西安
- 主办单位：中国通信学会
- 学会名称：中国通信学会
- 页数：6
- 文件大小：2805k
- 原文格式：O
- 会议级别：全国
摘要
运动船舶尺寸等参数的视频检测中,与船体同步运动的水面拖纹干扰会严重影响检测精度。为此,文中在描述显著性检测机理的基础上,提出了抑制船尾拖纹的船舶显著性视频检测算法:根据颜色对比度直方图得到内河场景的HC显著图;将原图超像素分割成若干子区域,以区域空间位置关系改进HC显著性检测结果得到区域显著图;通过该区域显著图初始化GrabCut算法,迭代分割过程中加入腐蚀膨胀操作来逼近目标边缘,从而提取运动船舶。经实况视频测试表明,该算法能有效地抑制船尾拖纹,准确地检测出内河运动船舶,准确性达到94.6%。
In the video detection of parameters such as ship size,the synchronous movement of the stern ripples will seriously affect the accuracy.Therefore,this paper proposes a novel algorithm of stern ripples-driven ship detection in inland waterway based on visual saliency.First,the algorithm utilizes a histogram-based contrast(HC) method to define HC saliency map for inland waterway using color statistics of the input image.Then,it performs super-pixel segmentation on original image to get several sub-regions and uses regional spatial relationship to improve HC saliency test results,which named regional saliency map.Finally,the algorithm performs initialization of GrabCut method with the regional saliency map,and has iterative process by adding erosion and dilation operations to get close to the target edge,so that the moving ship is extracted.Experimental results show that compared to other schemes the proposed approach can effectively restrain the stern ripple,accurately detect the ship in inland waterway,and its accuracy is up to 94.6%.
In the video detection of parameters such as ship size,the synchronous movement of the stern ripples will seriously affect the accuracy.Therefore,this paper proposes a novel algorithm of stern ripples-driven ship detection in inland waterway based on visual saliency.First,the algorithm utilizes a histogram-based contrast(HC) method to define HC saliency map for inland waterway using color statistics of the input image.Then,it performs super-pixel segmentation on original image to get several sub-regions and uses regional spatial relationship to improve HC saliency test results,which named regional saliency map.Finally,the algorithm performs initialization of GrabCut method with the regional saliency map,and has iterative process by adding erosion and dilation operations to get close to the target edge,so that the moving ship is extracted.Experimental results show that compared to other schemes the proposed approach can effectively restrain the stern ripple,accurately detect the ship in inland waterway,and its accuracy is up to 94.6%.
引文
[1]Teng Fi,Liu Qing.Muti-scale Ship racking via random projections[J].Signal.Image and Video Processing.2014,8(6):1069-1076.
[2]吕复强,王慧,刘泓.基于改进混合高斯模型的交通流背景检测算法IJ].上海交通大学学报.2012.46(12):1920-1925.Lv Fuqiang.Wang Hui.Liu Hong.Traffic Flow Background Detection Algorithm Based on Improved Gaussian Mixture Model[J].Journal of Shanghai Jiaotong Univerrsity.2012,46(12):1920-1925.
[3]Fefilatyev S,Goldgof D,and Lembke C.Tracking ships from fast moving camera through image registration[C]//Proceedings of the 2010 20th International Conference on Pattern Recognition,2010:3500-3503.
[4]Arshad N,Moon K,and Kim J.Multiple ship detection and tracking using background registration and morphological operations[C]//FGIT-SIP/MulGraB,2010:123,121-126,Springer Berlin Heidelberg.
[5]Li Yachao,Zhou Ruiyu,Quan Yinghui,Xing Mengdao.An Algorithm of Ship Target Detection Based on the Adaptive Background Window Function[J].Journal of Xi'an Jiaotong University,2013,47(6):25-30.
[6]Djelouah A,Franco J S,Boyer E,et al.Multi-view object segmentation in space and time[C]//2013IEEE International Conference on.IEEE,2013:628-635.
[7]訾玲玲,杜军中,汪谦.基于视觉显著性检测的跨尺度插值模型[J].北京邮电大学学报.2014,37(3):27-31.Zi Lingling,Du Junping,Wang Qian.Transcale Interpolation Model Based on Visual Saliency Detection[J].Journal of Beijing University of Posts and Telecommunications,2014,37(3):27-31.
[8]A Borji,D.N.Sihite,and L.Itti,What stands out in a scene?A study of human explicit saliency judgment[J].Vision research,2013,vol.91:62-77.
[9]Cheng Mingming,Niloy J.Mitra,Huang Xiaolei,Philip H.S.Torr,Hu Shimin.Global Contrast based Salient Region Detection[C]//IEEE Transactions on Partem Analysis and Machine Intelligence(IEEE TPAMI),2015.
[10]Ciomei I.Kyriakides E.Hybrid Ant Colony-Genetic Algorithm(GAA-PI)for Global Continuous Optimization[J].IEEE Transactions on Systems,Man and Cybernetics-Part B:Cybernetics,2012,42(1):234-245.
[11]C.Yang,L.Zhang,H.Lu,X.Ruan,and M.Yang.Saliency detection via graph-based manifold ranking[C]//Proceedings of IEEE CVPR,2013:3166-3173.
[12]Achanta R,Shaji A,Smith K,et al.SLIC.superpixels compared to state-of-the-art superpixel methods[J].Pattern Analysis and Machine Intelligence,IEEE Transactions on,2012,34(11):2274-2282.
[13]Rother C,Kolmogorov V,and Blake A."Grabcuf'-Interactive foreground extraction using iterated grap cuts[J].ACM Trans.Grap.2004,23(3):309-314.
[14]Christian Bontemps,Nour Meddahi.Testing distributional assumptions:A GMM approach[J].Journal of APPLIED ECONOMETRICS.2012,10:978-1012.
[15]胥欣,江登表,李勃,陈启美.混合高斯模型运动检测算法优化[J].计算机应用研究,2013,30(7):2190-2194.Xu Xin,Jiang Dengbiao,Li Bo,Chen Qimei.Optimization of Gaussian mixture model motion detection algorithm[J].Application Research of Computers,2013,30(7):2190-2194.
[16]Wijnhoven R,van Rens K,Jaspers EGT,and de With PHN.Online learning for ship detection in maritime surveillance[C]//Proc.the 32rd WIC Symposium on Infonnation Theory in the Benelux,2013,03:73-80.
[17]Luo Y,Yuan J,Xue P,et al.Saliency decsity maximization for efficient visual objects discovery[J].IEEE Transactions on Circuits and Systems for Video Technology,2011,21(12):1822-1834.
[2]吕复强,王慧,刘泓.基于改进混合高斯模型的交通流背景检测算法IJ].上海交通大学学报.2012.46(12):1920-1925.Lv Fuqiang.Wang Hui.Liu Hong.Traffic Flow Background Detection Algorithm Based on Improved Gaussian Mixture Model[J].Journal of Shanghai Jiaotong Univerrsity.2012,46(12):1920-1925.
[3]Fefilatyev S,Goldgof D,and Lembke C.Tracking ships from fast moving camera through image registration[C]//Proceedings of the 2010 20th International Conference on Pattern Recognition,2010:3500-3503.
[4]Arshad N,Moon K,and Kim J.Multiple ship detection and tracking using background registration and morphological operations[C]//FGIT-SIP/MulGraB,2010:123,121-126,Springer Berlin Heidelberg.
[5]Li Yachao,Zhou Ruiyu,Quan Yinghui,Xing Mengdao.An Algorithm of Ship Target Detection Based on the Adaptive Background Window Function[J].Journal of Xi'an Jiaotong University,2013,47(6):25-30.
[6]Djelouah A,Franco J S,Boyer E,et al.Multi-view object segmentation in space and time[C]//2013IEEE International Conference on.IEEE,2013:628-635.
[7]訾玲玲,杜军中,汪谦.基于视觉显著性检测的跨尺度插值模型[J].北京邮电大学学报.2014,37(3):27-31.Zi Lingling,Du Junping,Wang Qian.Transcale Interpolation Model Based on Visual Saliency Detection[J].Journal of Beijing University of Posts and Telecommunications,2014,37(3):27-31.
[8]A Borji,D.N.Sihite,and L.Itti,What stands out in a scene?A study of human explicit saliency judgment[J].Vision research,2013,vol.91:62-77.
[9]Cheng Mingming,Niloy J.Mitra,Huang Xiaolei,Philip H.S.Torr,Hu Shimin.Global Contrast based Salient Region Detection[C]//IEEE Transactions on Partem Analysis and Machine Intelligence(IEEE TPAMI),2015.
[10]Ciomei I.Kyriakides E.Hybrid Ant Colony-Genetic Algorithm(GAA-PI)for Global Continuous Optimization[J].IEEE Transactions on Systems,Man and Cybernetics-Part B:Cybernetics,2012,42(1):234-245.
[11]C.Yang,L.Zhang,H.Lu,X.Ruan,and M.Yang.Saliency detection via graph-based manifold ranking[C]//Proceedings of IEEE CVPR,2013:3166-3173.
[12]Achanta R,Shaji A,Smith K,et al.SLIC.superpixels compared to state-of-the-art superpixel methods[J].Pattern Analysis and Machine Intelligence,IEEE Transactions on,2012,34(11):2274-2282.
[13]Rother C,Kolmogorov V,and Blake A."Grabcuf'-Interactive foreground extraction using iterated grap cuts[J].ACM Trans.Grap.2004,23(3):309-314.
[14]Christian Bontemps,Nour Meddahi.Testing distributional assumptions:A GMM approach[J].Journal of APPLIED ECONOMETRICS.2012,10:978-1012.
[15]胥欣,江登表,李勃,陈启美.混合高斯模型运动检测算法优化[J].计算机应用研究,2013,30(7):2190-2194.Xu Xin,Jiang Dengbiao,Li Bo,Chen Qimei.Optimization of Gaussian mixture model motion detection algorithm[J].Application Research of Computers,2013,30(7):2190-2194.
[16]Wijnhoven R,van Rens K,Jaspers EGT,and de With PHN.Online learning for ship detection in maritime surveillance[C]//Proc.the 32rd WIC Symposium on Infonnation Theory in the Benelux,2013,03:73-80.
[17]Luo Y,Yuan J,Xue P,et al.Saliency decsity maximization for efficient visual objects discovery[J].IEEE Transactions on Circuits and Systems for Video Technology,2011,21(12):1822-1834.