-
针对某些光照变化、噪声不稳定等多模态场景不适合离线训练背景模型来提取目标信息的问题, 在基于混合高斯的背景建模的基础上, 利用帧间差分与邻域相似性实现模型初始参数的选取; 提出将随机子采样与邻域空间传播理论相结合改进参数更新过程; 在时间维度上建立观测向量, 实现模型参数的优化, 加快模型收敛速度; 并将颜色信息和梯度相融合实现基于多特征的多通道背景模型的建立, 采用背景点的随机采样策略简化多通道模型建立的计算量, 最终实现复杂环境下的运动目标的检测. 实验表明, 算法在抑制鬼影、动态背景和遮挡等方面有良好的检测性能, 且执行效率能够满足实时计算的需求.
In the process of tracking target, certain multi-modal background scenes are not suitable for the off-line training model, and moving target detection is affected as background in the current video environment is mostly multi-modal scene with much noise, and the characters of moving targets irregularly change, which,therefore, requires a more stable and robust moving target detection algorithm. To solve this problem, taking advantage of spatiotemporal relationship learning, the mixed Gaussian model (GMM) is improved in three aspects. First, the initialization method combining five-frame difference and intra-frame neighborhood average is proposed to obtain the initial parameters of the mixed Gaussian model. The five-frame difference method is introduced to obtain the initial parameters of the model, so that the background model is closer to the real scene. The intra-frame neighborhood average value is introduced, and an accumulation matrix CA is proposed to record the number of neighboring pixel points, then to enhance the information relevant to the neighborhood. This process can reduce the discontinuity of the target. Second, the calculation method of the neighborhood correlation is introduced to update the parameter of Gaussian model. Since the single pixel feature is related to the neighborhood random correlation, the random subsampling technology and neighborhood spatial propagation theory are combined together, and the execution efficiency is taken into account to simplify the process of updating model. To speed up the model convergence, an observation vector is built in the time dimension to optimize the model parameters, and the weight ω is gained based on the posterior probability. Then, the color-gradient method incooperated with the color HSI space and gradient information is adopted in this paper to complete the multi-channel Gaussian mixture model. The initial and the updated parameters of the Gaussian model in each channel can be acquired via the above steps. To simplify the computation of three channels, the random sampling of background pixels is introduced. Finally the detection of moving targets in complex environments is realized. The experiments show that the proposed algorithm has a great improvement in suppressing the influence of complex background and detecting target integrity, and the influence of the moving target in the initial stage is eliminated. -
Keywords:
- Gaussian mixture /
- random subsampling /
- neighborhood correlation /
- multi-channel
[1] Barron J L, Fleet D J, Beauchemin S S 1994 Computer Vision and Pattern Recognition Manufactured, Netherlands, February, 1994 p43
[2] 赖丽君, 徐智勇, 张栩铫 2016 红外与激光工程 45 273
Lai L J, Xu Z Y, Zhang X Y 2016 Infrared and Laser Engineering 45 273
[3] 崔智高, 王华, 李艾华 2017 物理学报 66 084203
Google Scholar
Cui Z G, Wang H, Li A H 2017 Acta Phys. Sin. 66 084203
Google Scholar
[4] Li W, Yao J G, Dong T Z, Li H 2016 International Congress on Image & Signal Processing Shenyang, China October 14—16, 2015 p969
[5] Staffer C, Grimson W L 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Fort Collins, Colorado, June 23—25, 1999 p2246
[6] Wren C R, Azarbayejani A, Darrell T, Pentland P A 1997 IEEE Transactions on Pattern Analysis and Machine Intelligence Washington, DC, USA 1997 p780
[7] 赵旭东, 刘鹏, 唐降龙 2011 自动化学报 37 915
Zhao X D, Liu P, Tang J L 2011 Acta Automatic Sinica 37 915
[8] Thierry B, Baf F E, Vachon B 2008 Recent Patents on Computer Science 1 219
Google Scholar
[9] Kaewtrakulpong P, Bowden R 2002 Kluwer Academic Publishers 135
[10] Kang K, Cao Y, Zhang J, Wang Z F 2016 Multimedia Tools & Applications 75 1443
[11] 李晓瑜, 马大中, 付英杰 2018 吉林大学学报(信息科学版) 36 61
Li X Y, Ma D Z, Fu Y J 2018 Journal of Jilin University (Information Science Edition)
36 61 [12] 朱文杰, 王广龙, 田杰, 乔中涛, 高凤岐 2018 北京理工大学学报 38 165
Zhu W J, Wang G L, Tian J, Qiao Z T, Gao F Q 2018 Journal of Beijing Institute of Technology 38 165
[13] Chen Y Y, Wang J Q, Lu H Q 2015 IEEE International Conference on Multimedia and Expo Turin, Italy, June 2 9
[14] Jeon M, Noh S, Noh S J, Jeon M 2012 Asian Conference on Computer Vision Daejeon, Korea, November 05—09, 2012 p493
[15] Choi M, Sweetman B 2013 Structural Health Monitoring 9 13
[16] Barnich O, Droogenbroeck M V 2011 IEEE Trans. Image Process. 20 1709
Google Scholar
[17] 徐艳, 董江涛, 王少华 2010 物理学报 59 7535
Google Scholar
Xu Y, Dong J T, Wang S H 2010 Acta Phys. Sin. 59 7535
Google Scholar
[18] Maha M A, Shedeed H A, Hussein A S 2010 International Conference on Image Processing Hong Kong September 26—29, 2010 p3453
[19] 李艳荻, 徐熙平, 陈江, 王鹤程 2017 仪器仪表学报 38 445
Li Y D, Xu X P, Chen J, Wang H C 2017 Chinese Journal of Scientific Instrument 38 445
[20] Martin D, Fahad S K, Michael F, Weijer J V D 2014 IEEE Conference on Computer Vision and Pattern Recognition Columbus, USA, June 23—28, 2014 p1090
[21] Jiang Y S, Ma J W 2015 IEEE Conference on Computer Vision and Pattern Recognition Boston, USA, June 7—12, 2015 p240
[22] Cuevas C, Yáñez E M, García N 2016 Comput. Vision and Image Understanding 152 p103
Google Scholar
[23] Goyette N, Pierre M J, Porikli F, Konrad J, Ishwar P 2012 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops Providence, Rhode Island, June 16—21, 2012 p1
-
图 4 算法精准度和准确率验证 (a)、(c) 精准度; (b)、(d) 准确率
Fig. 4. The accuracy and accuracy of the algorithm verify the results: (a)、(c) precision; (b)、(d) accuracy.
图 5 O_CL_01数据集第161帧的处理结果 (a) 原始图像; (b) 标注真值; (c) GMM; (d) RGB_GMM; (e) ViBe; (f) HSG_GMM
Fig. 5. The result of the frame 161 in O_CL_01 data set: (a) Original image; (b) true Value image; (c) GMM result; (d) RGB_GMM result; (e) ViBe result; (f) HSG_GMM result.
图 6 初始视频中存在动目标的静态背景检验结果 (a)原始图像; (b)GMM算法结果; (c) RGB-GMM结果; (d) ViBe结果; (e) HSG-GMM结果
Fig. 6. Detection result of moving target in initial video in static background: (a) Initial image; (b) GMM result; (c) RGB-GMM result; (d) ViBe result; (e) HSG-GMM result.
-
[1] Barron J L, Fleet D J, Beauchemin S S 1994 Computer Vision and Pattern Recognition Manufactured, Netherlands, February, 1994 p43
[2] 赖丽君, 徐智勇, 张栩铫 2016 红外与激光工程 45 273
Lai L J, Xu Z Y, Zhang X Y 2016 Infrared and Laser Engineering 45 273
[3] 崔智高, 王华, 李艾华 2017 物理学报 66 084203
Google Scholar
Cui Z G, Wang H, Li A H 2017 Acta Phys. Sin. 66 084203
Google Scholar
[4] Li W, Yao J G, Dong T Z, Li H 2016 International Congress on Image & Signal Processing Shenyang, China October 14—16, 2015 p969
[5] Staffer C, Grimson W L 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Fort Collins, Colorado, June 23—25, 1999 p2246
[6] Wren C R, Azarbayejani A, Darrell T, Pentland P A 1997 IEEE Transactions on Pattern Analysis and Machine Intelligence Washington, DC, USA 1997 p780
[7] 赵旭东, 刘鹏, 唐降龙 2011 自动化学报 37 915
Zhao X D, Liu P, Tang J L 2011 Acta Automatic Sinica 37 915
[8] Thierry B, Baf F E, Vachon B 2008 Recent Patents on Computer Science 1 219
Google Scholar
[9] Kaewtrakulpong P, Bowden R 2002 Kluwer Academic Publishers 135
[10] Kang K, Cao Y, Zhang J, Wang Z F 2016 Multimedia Tools & Applications 75 1443
[11] 李晓瑜, 马大中, 付英杰 2018 吉林大学学报(信息科学版) 36 61
Li X Y, Ma D Z, Fu Y J 2018 Journal of Jilin University (Information Science Edition)
36 61 [12] 朱文杰, 王广龙, 田杰, 乔中涛, 高凤岐 2018 北京理工大学学报 38 165
Zhu W J, Wang G L, Tian J, Qiao Z T, Gao F Q 2018 Journal of Beijing Institute of Technology 38 165
[13] Chen Y Y, Wang J Q, Lu H Q 2015 IEEE International Conference on Multimedia and Expo Turin, Italy, June 2 9
[14] Jeon M, Noh S, Noh S J, Jeon M 2012 Asian Conference on Computer Vision Daejeon, Korea, November 05—09, 2012 p493
[15] Choi M, Sweetman B 2013 Structural Health Monitoring 9 13
[16] Barnich O, Droogenbroeck M V 2011 IEEE Trans. Image Process. 20 1709
Google Scholar
[17] 徐艳, 董江涛, 王少华 2010 物理学报 59 7535
Google Scholar
Xu Y, Dong J T, Wang S H 2010 Acta Phys. Sin. 59 7535
Google Scholar
[18] Maha M A, Shedeed H A, Hussein A S 2010 International Conference on Image Processing Hong Kong September 26—29, 2010 p3453
[19] 李艳荻, 徐熙平, 陈江, 王鹤程 2017 仪器仪表学报 38 445
Li Y D, Xu X P, Chen J, Wang H C 2017 Chinese Journal of Scientific Instrument 38 445
[20] Martin D, Fahad S K, Michael F, Weijer J V D 2014 IEEE Conference on Computer Vision and Pattern Recognition Columbus, USA, June 23—28, 2014 p1090
[21] Jiang Y S, Ma J W 2015 IEEE Conference on Computer Vision and Pattern Recognition Boston, USA, June 7—12, 2015 p240
[22] Cuevas C, Yáñez E M, García N 2016 Comput. Vision and Image Understanding 152 p103
Google Scholar
[23] Goyette N, Pierre M J, Porikli F, Konrad J, Ishwar P 2012 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops Providence, Rhode Island, June 16—21, 2012 p1
下载:
计量
- 文章访问数: 7069
- PDF下载量: 45
- 被引次数: 0
