一种新的矿井监控视频增强目标检测算法
|
摘要
由于矿井下光线不足,照度低且粉尘大,造成监控视频图像存在昏暗和模糊问题,利用小波变换获取视频画面中的不同频率分量信息,首先对低频分量采用暗原色先验进行去雾处理,然后用阈值滤波对高频分量进行消噪,将处理以后的低频分量和高频分量进行融合,重构小波函数,实现图像的增强。仿真实验结果表明所提算法能提高图像对比度,增强图像细节信息,淡化浓雾、抑制噪声等方面有较好的效果。在矿井运动目标检测中,为了改善传统混合高斯模型像素点不能精确匹配及参数迭代速度慢的问题,采用三帧差分法融合混合高斯背景模型,融合后的算法有效消除了背景更新不及时而导致的画面鬼影现象,而且运算速度得到明显提升,实现了运动目标实时追踪的需求。仿真实验结果表明所提算法相对传统混合高斯模型算法不仅能够快速的检测出运动目标,而且检测图像边缘细节信息更加清晰,并且能够解决物体遮挡等问题,为矿井视频信息处理和人员安全监测奠定了良好的基础。
Aiming at the problems of dim lighting,low illumination and large dust in underground mine,which would lead to the dim and fuzzy of video,wavelet transformation is used to get the component of different frequency. The low-frequency component is firstly defogged using dark primary priori,and highfrequency component is denoised by threshold filter. The processed low-frequency component and highfrequency component are fused to reconstruct the wavelet function to realize the image enhancement. The simulation results show that the proposed algorithm can improve image contrast,enhance image details,reduce fog and suppress noise. Aiming at the problem that the traditional Gaussian mixture model can neither accurately match the pixels nor update the parameters fast in moving target detection,the threeframe difference method is used to fuse the Gaussian mixture background model. The fused algorithm eliminates the ghost phenomenon of the Gaussian background model caused by the untimely updating of the background,and improves the operation speed of the algorithm,meeting the real-time requirement.The results show that the proposed algorithm can not only detect moving targets faster than the traditional Mixture Gauss Model algorithm,but also detect image edge details more clearly,and can solve the problem of object occlusion,which lays a good foundation for mine video information processing and personnel safety monitoring.
Aiming at the problems of dim lighting,low illumination and large dust in underground mine,which would lead to the dim and fuzzy of video,wavelet transformation is used to get the component of different frequency. The low-frequency component is firstly defogged using dark primary priori,and highfrequency component is denoised by threshold filter. The processed low-frequency component and highfrequency component are fused to reconstruct the wavelet function to realize the image enhancement. The simulation results show that the proposed algorithm can improve image contrast,enhance image details,reduce fog and suppress noise. Aiming at the problem that the traditional Gaussian mixture model can neither accurately match the pixels nor update the parameters fast in moving target detection,the threeframe difference method is used to fuse the Gaussian mixture background model. The fused algorithm eliminates the ghost phenomenon of the Gaussian background model caused by the untimely updating of the background,and improves the operation speed of the algorithm,meeting the real-time requirement.The results show that the proposed algorithm can not only detect moving targets faster than the traditional Mixture Gauss Model algorithm,but also detect image edge details more clearly,and can solve the problem of object occlusion,which lays a good foundation for mine video information processing and personnel safety monitoring.
引文
[1]李利华,栾晓峰.基于空域、频域和时间域的复合图像增强方法[J].医疗装备,2018,31(1):45-46.LI Li-hua,LUAN Xiao-feng. Composite image enhancement method based on airspace,frequency domain and time domain[J]. Medical Equipment,2018,31(1):45-46.
[2]赵翱东,奚茂龙,叶茜.基于奇异值分解和非下采样Contourlet变换的红外图像增强新算法[J].山东农业大学学报,2018,49(5):852-855.ZHAO Ao-dong,XI Mao-long,YE Qian. A new infrared image enhancement algorithm based on singular value decomposition and nonsubsampled contourlet transform[J]. Journal of Shandong Agricultural University,2018,49(5):852-855.
[3]高银,云利军,石俊生.基于四阶PDE模型的暗原色理论雾天图像增强算法[J].南京理工大学学报,2015,39(5):544-549.GAO Yin,YUN Li-jun,SHI Jun-sheng. A fog-day image enhancement algorithm based on fourth-order PDE model for dark primary color theory[J]. Journal of Nanjing University of Science and Technology,2015,39(5):544-549.
[4]卢光森,李莹,马敏.基于新阈值函数的小波阈值去噪算法[J].传感器与微系统,2017,36(12):141-144.LU Guang-sen,LI Ying,MA Min. Wavelet threshold denoising algorithm based on new threshold function[J].Transducer and Microsystem,2017,36(12):141-144.
[5]赵洪图,刘云.基于三次样条插值的小波模极大值去噪算法[J].计算机工程与设计,2014,35(8):2965-2968.ZHAO Hong-tu,LIU Yun. A wavelet modulus maximum denoising algorithm based on cubic spline interpolation[J]. Computer Engineering and Design,2014,35(8):2965-2968.
[6]夏欣雨,朱春梅,吴国新.融合小波变换和模糊集的东巴古籍图像增强算法[J].北京信息科技大学学报,2018,33(3):47-51,56.XIA Xin-yu,ZHU Chun-mei,WU Guo-xin. Image enhancement algorithm of Dongba ancient books based on wavelet transform and fuzzy set[J]. Journal of Beijing Information Science and Technology University,2018,33(3):47-51,56.
[7]于雯越,安博文,赵明.基于光流法与RPCA的红外运动目标检测[J].现代计算机,2018(23):66-71.YU Wen-yue,AN Bo-wen,ZHAO Ming. Infrared moving target detection based on optical flow method and RPCA[J]. Modern Computer,2018(23):66-71.
[8]范长军,文凌艳,毛泉涌,等.结合单高斯与光流法的无人机运行目标检测[J].计算机系统应用,2019,28(2):184-189.FAN Chang-jun,WEN Ling-yan,MAO Quan-yong,et al.UAV moving target detection based on single Gauss and optical flow method[J]. Computer System Application,2019,28(2):184-189.
[9]伍健荣,杜向龙,刘海涛.一种改进的基于卡尔曼滤波的自适应背景建模算法[J].传感器与微系统技术,2012,31(1):52-54,58.WU Jian-rong,DU Xiang-long,LIU Hai-tao. An improved adaptive background modeling algorithm based on Kalman filter[J]. Transducer&Micro System Technologies,2012,31(1):52-54,58.
[10] Thierry Bouwmans,Caroline Silva,Cristina Marghes,et al. On the role and the importance of features for background modeling and foreground detection[J]. Computer Science Review,2018,28:26-91.
[11]严丽罗,韩林万,李夏雪.基于混合背景减法和帧差的自适应运动车辆检测算法[J].应用力学与材料,2014,3468(644-650):930-933.YAN Li-luo,HAN Lin-wan,LI Xia-xue. Adaptive moving vehicle detection algorithm based on hybrid background subtraction and frame difference[J]. Applied Mechanics and Materials,2014,3468(644-650):930-933.
[12] Ramya P,Rajeswari R. A modified frame difference method using correlation coefficient for background subtraction[J]. Procedia Computer Science,2016,93:478-485.
[13] LU Gui-liang,ZHOU Yu,YU Yao-deng. A novel approach for foreign substances detection in injection using clustering and frame difference[J]. Sensors,2011,11(10):9121-9135.
[14]田小平,乔森,吴成茂.一种改进的图像模糊对比度增强算法[J].西安邮电学院学报,2017,22(3):50-54.TIAN Xiao-ping,QIAO Sen,WU Cheng-mao. An improved algorithm for image fuzzy contrast enhancement[J]. Journal of Xi’an University of Posts&Telecommunications,2017,22(3):50-54.
[15]李晓瑜,马大中,付英杰.基于三帧差分混合高斯背景模型运动目标检测[J].吉林大学学报,2018,36(4):414-422.LI Xiao-yu,MA Da-zhong,FU Ying-jie. Moving target detection based on three-frame differential mixed Gaussian background model[J]. Journal of Jilin University,2018,36(4):414-422.
[16]施龙超,安玉磊,苏秉华,等.一种改进的基于卡尔曼滤波的背景差分算法[J].激光与光电子学进展,2018,55(8):216-222.SHI Long-chao,AN Yu-lei,SU Bing-hua,et al. An improved background difference algorithm based on Kalman filter[J]. Laser&Optoelectronics Progress,2018,55(8):216-222.
[17] Christ of Ridder,Olaf Munkelt,Harald Kirchner.基于卡尔曼滤波的自适应背景估计与前景检测[C]//最近关于机电一体化进展的国际会议记录,2015:193-199.Christ of Ridder,Olaf Munkelt,Harald Kirchner. Adaptive background estimation and foreground detection using Kalman-Filtering[C]//Proceedings of International Conference on recent Advances in Mechatronics,2015:193-199.
[18]黄大卫,胡文翔,吴小培,等.改进单高斯模型的视频前景提取与破碎目标合并算法[J].信号处理,2015,31(3):299-307.HUANG Da-wei,HU Wen-xiang,WU Xiao-pei,et al.Video foreground extraction and fragmentation target merging algorithm for improved single Gaussian model[J]. Signal Processing,2015,31(3):299-307.
[19]张燕平,白云球,赵勇,等.应用改进混合高斯模型的运动目标检测[J].计算机工程与应用,2010,46(34):155-157,223.ZHANG Yan-ping,BAI Yun-qiu,ZHAO Yong,et al.Moving object detection using improved Gaussian model[J]. Computer Engineering and Applications,2010,46(34):155-157,223.
[20]汪成亮,周佳,黄晟.基于高斯混合模型与PCAHOG的快速运动人体检测[J].计算机应用研究,2012,29(6):2156-2160.WANG Cheng-liang,ZHOU Jia,HUANG Sheng. Fast moving human detection based on Gaussian mixture model and PCA-HOG[J]. Application Research of Computers,2012,29(6):2156-2160.
[21]习通.基于稀疏光流检测的目标跟踪方法研究[J].科技风,2018(6):207.XI Tong. Research on target tracking method based on sparse optical flow detection[J]. Science and Technology,2018(6):207.
[22]周莉鸿.改进的混合高斯模型运动目标检测算法[J].电子科技,2017,30(7):21-24.ZHOU Li-hong. Improved hybrid Gaussian model moving target detection algorithm[J]. Electronic Technology,2017,30(7):21-24.
[2]赵翱东,奚茂龙,叶茜.基于奇异值分解和非下采样Contourlet变换的红外图像增强新算法[J].山东农业大学学报,2018,49(5):852-855.ZHAO Ao-dong,XI Mao-long,YE Qian. A new infrared image enhancement algorithm based on singular value decomposition and nonsubsampled contourlet transform[J]. Journal of Shandong Agricultural University,2018,49(5):852-855.
[3]高银,云利军,石俊生.基于四阶PDE模型的暗原色理论雾天图像增强算法[J].南京理工大学学报,2015,39(5):544-549.GAO Yin,YUN Li-jun,SHI Jun-sheng. A fog-day image enhancement algorithm based on fourth-order PDE model for dark primary color theory[J]. Journal of Nanjing University of Science and Technology,2015,39(5):544-549.
[4]卢光森,李莹,马敏.基于新阈值函数的小波阈值去噪算法[J].传感器与微系统,2017,36(12):141-144.LU Guang-sen,LI Ying,MA Min. Wavelet threshold denoising algorithm based on new threshold function[J].Transducer and Microsystem,2017,36(12):141-144.
[5]赵洪图,刘云.基于三次样条插值的小波模极大值去噪算法[J].计算机工程与设计,2014,35(8):2965-2968.ZHAO Hong-tu,LIU Yun. A wavelet modulus maximum denoising algorithm based on cubic spline interpolation[J]. Computer Engineering and Design,2014,35(8):2965-2968.
[6]夏欣雨,朱春梅,吴国新.融合小波变换和模糊集的东巴古籍图像增强算法[J].北京信息科技大学学报,2018,33(3):47-51,56.XIA Xin-yu,ZHU Chun-mei,WU Guo-xin. Image enhancement algorithm of Dongba ancient books based on wavelet transform and fuzzy set[J]. Journal of Beijing Information Science and Technology University,2018,33(3):47-51,56.
[7]于雯越,安博文,赵明.基于光流法与RPCA的红外运动目标检测[J].现代计算机,2018(23):66-71.YU Wen-yue,AN Bo-wen,ZHAO Ming. Infrared moving target detection based on optical flow method and RPCA[J]. Modern Computer,2018(23):66-71.
[8]范长军,文凌艳,毛泉涌,等.结合单高斯与光流法的无人机运行目标检测[J].计算机系统应用,2019,28(2):184-189.FAN Chang-jun,WEN Ling-yan,MAO Quan-yong,et al.UAV moving target detection based on single Gauss and optical flow method[J]. Computer System Application,2019,28(2):184-189.
[9]伍健荣,杜向龙,刘海涛.一种改进的基于卡尔曼滤波的自适应背景建模算法[J].传感器与微系统技术,2012,31(1):52-54,58.WU Jian-rong,DU Xiang-long,LIU Hai-tao. An improved adaptive background modeling algorithm based on Kalman filter[J]. Transducer&Micro System Technologies,2012,31(1):52-54,58.
[10] Thierry Bouwmans,Caroline Silva,Cristina Marghes,et al. On the role and the importance of features for background modeling and foreground detection[J]. Computer Science Review,2018,28:26-91.
[11]严丽罗,韩林万,李夏雪.基于混合背景减法和帧差的自适应运动车辆检测算法[J].应用力学与材料,2014,3468(644-650):930-933.YAN Li-luo,HAN Lin-wan,LI Xia-xue. Adaptive moving vehicle detection algorithm based on hybrid background subtraction and frame difference[J]. Applied Mechanics and Materials,2014,3468(644-650):930-933.
[12] Ramya P,Rajeswari R. A modified frame difference method using correlation coefficient for background subtraction[J]. Procedia Computer Science,2016,93:478-485.
[13] LU Gui-liang,ZHOU Yu,YU Yao-deng. A novel approach for foreign substances detection in injection using clustering and frame difference[J]. Sensors,2011,11(10):9121-9135.
[14]田小平,乔森,吴成茂.一种改进的图像模糊对比度增强算法[J].西安邮电学院学报,2017,22(3):50-54.TIAN Xiao-ping,QIAO Sen,WU Cheng-mao. An improved algorithm for image fuzzy contrast enhancement[J]. Journal of Xi’an University of Posts&Telecommunications,2017,22(3):50-54.
[15]李晓瑜,马大中,付英杰.基于三帧差分混合高斯背景模型运动目标检测[J].吉林大学学报,2018,36(4):414-422.LI Xiao-yu,MA Da-zhong,FU Ying-jie. Moving target detection based on three-frame differential mixed Gaussian background model[J]. Journal of Jilin University,2018,36(4):414-422.
[16]施龙超,安玉磊,苏秉华,等.一种改进的基于卡尔曼滤波的背景差分算法[J].激光与光电子学进展,2018,55(8):216-222.SHI Long-chao,AN Yu-lei,SU Bing-hua,et al. An improved background difference algorithm based on Kalman filter[J]. Laser&Optoelectronics Progress,2018,55(8):216-222.
[17] Christ of Ridder,Olaf Munkelt,Harald Kirchner.基于卡尔曼滤波的自适应背景估计与前景检测[C]//最近关于机电一体化进展的国际会议记录,2015:193-199.Christ of Ridder,Olaf Munkelt,Harald Kirchner. Adaptive background estimation and foreground detection using Kalman-Filtering[C]//Proceedings of International Conference on recent Advances in Mechatronics,2015:193-199.
[18]黄大卫,胡文翔,吴小培,等.改进单高斯模型的视频前景提取与破碎目标合并算法[J].信号处理,2015,31(3):299-307.HUANG Da-wei,HU Wen-xiang,WU Xiao-pei,et al.Video foreground extraction and fragmentation target merging algorithm for improved single Gaussian model[J]. Signal Processing,2015,31(3):299-307.
[19]张燕平,白云球,赵勇,等.应用改进混合高斯模型的运动目标检测[J].计算机工程与应用,2010,46(34):155-157,223.ZHANG Yan-ping,BAI Yun-qiu,ZHAO Yong,et al.Moving object detection using improved Gaussian model[J]. Computer Engineering and Applications,2010,46(34):155-157,223.
[20]汪成亮,周佳,黄晟.基于高斯混合模型与PCAHOG的快速运动人体检测[J].计算机应用研究,2012,29(6):2156-2160.WANG Cheng-liang,ZHOU Jia,HUANG Sheng. Fast moving human detection based on Gaussian mixture model and PCA-HOG[J]. Application Research of Computers,2012,29(6):2156-2160.
[21]习通.基于稀疏光流检测的目标跟踪方法研究[J].科技风,2018(6):207.XI Tong. Research on target tracking method based on sparse optical flow detection[J]. Science and Technology,2018(6):207.
[22]周莉鸿.改进的混合高斯模型运动目标检测算法[J].电子科技,2017,30(7):21-24.ZHOU Li-hong. Improved hybrid Gaussian model moving target detection algorithm[J]. Electronic Technology,2017,30(7):21-24.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |