基于全景视觉的行人检测技术研究
|
摘要
行人检测是当前计算机视觉和人工智能领域研究的重点和热点,也是目标检测的重要分支,其在军事领域、智能交通、机器人导航、智能监控、人体运动分析等领域都有广泛的应用前景。由于人体的衣着、姿势、光照各异,行人检测是非常具有挑战性的课题。目前,计算机视觉领域的主流思想将行人检测问题认为是一个模式分类问题,其方法是从大量的行人训练样本中提取特征,利用机器学习方法将行人与其他运动目标以及干扰背景区分开来,并准确定位。全景视觉系统具有视场范围大的特点,当前已经广泛地应用在机器人导航、空间探测、视频监控、虚拟现实、环境感知技术等领域。行人检测技术和全景视觉系统的结合更能充分发挥各自的优势,提供更加广泛的应用前景。在全景视觉中研究行人检测,可以充分利用全景视觉视野范围大的特点,检测到360°视野范围内的行人,更好的满足机器人导航和智能监控等应用场合的需求。近几年来,GPU和基于GPU的通用计算得到了迅速的发展,基于GPU的通用计算在计算机视觉领域得到广泛的应用。
本文主要研究基于双曲面折反射全景视觉系统的行人检测方法,采用积分梯度方向直方图特征和线性支持向量机设计了基于全景视觉的行人检测器,并在GPU平台上利用并行计算技术改进了行人检测算法的关键步骤,在保持检测准确率的前提下,获得明显的检测速度上的提升。
首先研究了双曲面折反射全景视觉的成像机理,推导了双曲面折反射成像系统的数学模型。为了克服全景图像畸变对行人检测的影响,对原始全景图像进行还原解算,从而获得符合正常人的视觉感受的全景图像,保证全景视觉中行人检测的准确性。
然后本文设计了一个基于全景视觉的行人检测器,采用Dalal提出的梯度方向直方图特征为基础,应用了用积分图像方法来计算HOG特征集的方法,大大提高了HOG特征的计算速度,同时采用线性支持向量机作为行人分类器,对在行人检测中存在大量的重叠结果窗口,采用非极大值抑制算法进行多结果融合,获得较好的检测效果。
最后本文在GPU平台上对基于全景视觉的行人检测器进行了改进,采用CUDA并行计算技术加速全景图像还原解算,积分梯度方向直方图计算和线性支持向量机检测。本文对算法的关键步骤所需时间以及资源占用情况进行了详细的分析。实验结果表明,该方法在保持行人检测准确性的同时,极大地提升了在全景图像中行人检测的速度,我们的算法所需时间只是在CPU上实现的算法的1/8左右。
Nowadays, pedestrian detection is intensively investigated and becoming a hot topic in the fields of computer vision and artificial intelligence. It's also an important branch of object detection. It could be widely used in the military field,intelligent transportation systems,robot navigation, intelligent surveillance, human motion analysis and so on. Detecting human in images is a challenging task because of the variability in clothing and illumination conditions, and the wide range of poses that people can adopt。Nowadays, pedestrian detection is considered as a pattern classification problem in the filed of computer vision. The Algorithm mainly based on machine learning, which extracts features from pedestrian samples and distinguish pedestrian from other targets and background area, to find the accurate location of pedestrian.Panoramic vision system has a large field of view.It's widely used in robot navigation, space probe,video surveillance, virtual reality, environment sensing technology and so on. The combination of pedestrian detection technology and panoramic vision system could play their respective advantages.It's potential application is very promising.Pedestrian detection based on panoramic vision system could take advantage of a large panoramic field of view and detect pedestrian in the 360-degree field of view, to better meet needs of robot navigation intelligent surveillance and other applications. Last few years, Graphics processing unit and general-purpose computing on GPU has been in rapid development,general-purpose computing on graphics processing units has been widely used in computer vision.
This paper mainly studies the pedestrian detection algorithm based on panoramic vision system, designs a pedestrian detector based on panoramic vision with integral histograms of oriented gradient features and linear support vector machine, and improves the key steps of the pedestrian detection algorithm with parallel computing technology on GPU platform, achieves significant improvement on the detection speed, while maintaining the premise of detection accuracy。
Firstly, by researching the imaging principle of panoramic vision, derive mathematical model of the hyperboloid reflection imaging system. In order to overcome the impact of panoramic image distortion on pedestrian detection,a cylinder unwarping algorithm based on geometry principle of panoramic image is proposed can obtain 360-dcgrce panoramic image. Ensure the detection accuracy of pedestrian detection in panoramic image.
Secondly, this paper designs a pedestrian detector based on panoramic vision, based on histogram of oriented gradient features proposed by Dalal, Introduced with the integral image method to calculate the HOG feature set, greatly improved the calculation speed of HOG features. At the same time using a linear support vector machine as a pedestrian classifier. In the pedestrian detection, there are a large number of overlapping windows, we use non-maxima suppression algorithm for integration of multiple results,to obtain better detection results.
Finally, this paper improved the pedestrian detector based on panoramic vision on GPU platform,using CUDA parallel computing technology to accelerate to obtain panoramic image, calculate the integral histogram of oriented gradient and the linear support vector machine detection. we detailed analyzes the processing times and the occupancy of each kernel used by our algorithm. Experimental results show that our method achieves significant improvement on the detection speed in panoramic image, and we report a speed up by a factor of 8 over our CPU implementation, while maintaining the premise of detection accuracy.
本文主要研究基于双曲面折反射全景视觉系统的行人检测方法,采用积分梯度方向直方图特征和线性支持向量机设计了基于全景视觉的行人检测器,并在GPU平台上利用并行计算技术改进了行人检测算法的关键步骤,在保持检测准确率的前提下,获得明显的检测速度上的提升。
首先研究了双曲面折反射全景视觉的成像机理,推导了双曲面折反射成像系统的数学模型。为了克服全景图像畸变对行人检测的影响,对原始全景图像进行还原解算,从而获得符合正常人的视觉感受的全景图像,保证全景视觉中行人检测的准确性。
然后本文设计了一个基于全景视觉的行人检测器,采用Dalal提出的梯度方向直方图特征为基础,应用了用积分图像方法来计算HOG特征集的方法,大大提高了HOG特征的计算速度,同时采用线性支持向量机作为行人分类器,对在行人检测中存在大量的重叠结果窗口,采用非极大值抑制算法进行多结果融合,获得较好的检测效果。
最后本文在GPU平台上对基于全景视觉的行人检测器进行了改进,采用CUDA并行计算技术加速全景图像还原解算,积分梯度方向直方图计算和线性支持向量机检测。本文对算法的关键步骤所需时间以及资源占用情况进行了详细的分析。实验结果表明,该方法在保持行人检测准确性的同时,极大地提升了在全景图像中行人检测的速度,我们的算法所需时间只是在CPU上实现的算法的1/8左右。
Nowadays, pedestrian detection is intensively investigated and becoming a hot topic in the fields of computer vision and artificial intelligence. It's also an important branch of object detection. It could be widely used in the military field,intelligent transportation systems,robot navigation, intelligent surveillance, human motion analysis and so on. Detecting human in images is a challenging task because of the variability in clothing and illumination conditions, and the wide range of poses that people can adopt。Nowadays, pedestrian detection is considered as a pattern classification problem in the filed of computer vision. The Algorithm mainly based on machine learning, which extracts features from pedestrian samples and distinguish pedestrian from other targets and background area, to find the accurate location of pedestrian.Panoramic vision system has a large field of view.It's widely used in robot navigation, space probe,video surveillance, virtual reality, environment sensing technology and so on. The combination of pedestrian detection technology and panoramic vision system could play their respective advantages.It's potential application is very promising.Pedestrian detection based on panoramic vision system could take advantage of a large panoramic field of view and detect pedestrian in the 360-degree field of view, to better meet needs of robot navigation intelligent surveillance and other applications. Last few years, Graphics processing unit and general-purpose computing on GPU has been in rapid development,general-purpose computing on graphics processing units has been widely used in computer vision.
This paper mainly studies the pedestrian detection algorithm based on panoramic vision system, designs a pedestrian detector based on panoramic vision with integral histograms of oriented gradient features and linear support vector machine, and improves the key steps of the pedestrian detection algorithm with parallel computing technology on GPU platform, achieves significant improvement on the detection speed, while maintaining the premise of detection accuracy。
Firstly, by researching the imaging principle of panoramic vision, derive mathematical model of the hyperboloid reflection imaging system. In order to overcome the impact of panoramic image distortion on pedestrian detection,a cylinder unwarping algorithm based on geometry principle of panoramic image is proposed can obtain 360-dcgrce panoramic image. Ensure the detection accuracy of pedestrian detection in panoramic image.
Secondly, this paper designs a pedestrian detector based on panoramic vision, based on histogram of oriented gradient features proposed by Dalal, Introduced with the integral image method to calculate the HOG feature set, greatly improved the calculation speed of HOG features. At the same time using a linear support vector machine as a pedestrian classifier. In the pedestrian detection, there are a large number of overlapping windows, we use non-maxima suppression algorithm for integration of multiple results,to obtain better detection results.
Finally, this paper improved the pedestrian detector based on panoramic vision on GPU platform,using CUDA parallel computing technology to accelerate to obtain panoramic image, calculate the integral histogram of oriented gradient and the linear support vector machine detection. we detailed analyzes the processing times and the occupancy of each kernel used by our algorithm. Experimental results show that our method achieves significant improvement on the detection speed in panoramic image, and we report a speed up by a factor of 8 over our CPU implementation, while maintaining the premise of detection accuracy.
引文
[1]David Wettergreen, Operating Nomad during the Atacama Desert Trek. Proceedings of IEEE ICRA,1998:1016-1023P
[2]Paul E. Rybski, Fernando de la Torre, Raju Patil, Carlos Vallespi, Manuela M. Veloso, Brett Browning. CAMEO:Camera Assisted Meeting Event Observer. Robotics and Automation. Proceedings of IEEE ICRA,2004:1634-1639P
[3]Patil R, Rybski P E, Kanade T. People detection and tracking in high resolution panoramic video mosaic. Intelligent Robots and Systems, Proceedings.of IEEE/RSJ. 2004:1323-1328P
[4]Chakravarty P, Jarvis R. Panoramic Vision and Laser Range Finder Fusion for Multiple Person Tracking. Intelligent Robots and Systems, Proceedings of IEEE/RSJ. 2006:2949-2954P
[5]Koyasu H, Miura J, Shirai Y. Real-time omnidirectional stereo for obstacle detection and tracking in dynamic environments. Intelligent Robots and Systems, Proceedings of IEEE/RSJ.2001:31-36P
[6]Baba A, Chatila R. Dynamic Targets Detection for Robotic Applications Using Panoramic Vision System. Recent Progress in Robotics:Viable Robotic Service to Human.2008:215-227P
[7]曾吉勇苏显渝.双曲面折反射全景成像系统.光学学报,2003(9):1138-1142页
[8]曾吉勇苏显渝.折反射全景成像系统.激光杂志,2004(6):62-64页
[9]魏芳,董再励,孙茂相,王晓蕾.用于移动机器人的视觉全局定位系统研究.机器人,2001(5)：400-403页
[10]张尧陈卫东.一个基于全景视觉的移动机器人导航系统的设计与实现.机器人,2005(2)：173-177页
[11]王景川陈卫东,曹其新.基于全景视觉与里程计的移动机器人自定位方法研究.机器人,2005(1)：41-45页
[12]白剑,牛爽,杨国光.全景光学环带凝视成像技术.红外与激光工程,2006(3):331-335页
[13]肖潇,杨国光.全景成像技术的现状和进展.光学仪器,2007(4)：84-89页
[14]肖潇,杨国光,白剑.基于球面透视投影约束的全景环形透镜畸变校正.光学学报,2008(4)：675-680页
[15]席志红.全方位视觉技术及其在智能移动机器人等领域的应用研究.哈尔滨工程大学博士学位论文,2006:12-16页
[16]夏桂华,王博,朱齐丹.改进SIFT用于全景视觉移动机器人定位.计算机工程与应用,2010(18)：196-198页.
[17]Viola P, Jones M J, Snow D. Detecting pedestrians using patterns of motion and appearance. Computer Vision,2003:734-741P
[18]N.Dalal and B.Triggs Histograms of oriented gradients for human detection. SanDiego, CA, UnitedStates,2005.Institute of Electrical and Eleetronics Engineers Computer Soeiety, Piscataway, NJ 08855-1331, UnitedStates,886-893P
[19]Qiang Zhu, Mei-Chen Yeh, Kwang-Ting Cheng. Fast Human Detection Using a Cascade of Histograms of Oriented Gradients. Computer Vision and Pattern Recognition,2006:1491-1498P
[20]He X, Li J, Chen Y. Local Binary Patterns with Mahalanobis Distance Maps for Human Detection. Image and Signal Processing.2008:520-524P
[21]Shet V D, Neumann J, Ramesh V. Bilattice-based Logical Reasoning for Human Detection. Computer Vision and Pattern Recognition.2007:1-8P
[22]Ying Wu, Ting Yu. A field model for human detection and tracking. Pattern Analysis and Machine Intelligence,2006,28(5):753-765P
[23]Ying Wu, Ting Yu, Gang Hua. A statistical field model for pedestrian detection. Computer Vision and Pattern Recognition,2005:1023-1030P
[24]Zeng C, Ma H, Ming A. Fast human detection using mi-sVM and a cascade of HOG-LBP features. Image Processing (ICIP),2010:3845-3848P
[25]李同治,丁晓青,王生进.利用级联SVM的人体检测方法.中国图象图形学报,2008(3)：566-570页
[26]贾慧星,章毓晋.车辆辅助驾驶系统中基于计算机视觉的行人检测研究综述[J.自动化学报,2007(1)：84-90页
[27]胡斌,王生进,丁晓青.基于部位检测和子结构组合的行人检测方法.计算机科学,2009(11)：242-246页
[28]许言午.面向行人检测的组合分类计算模型与应用研究.中国科学技术大学博士学位论文,2009:26-30页
[29]杨志辉.基于多尺度方向特征的行人检测算法.哈尔滨工程大学硕士学位论文,2009：36-41页
[30]吴贻军.基于视频序列的运动人体检测算法研究.浙江大学硕士学位论文,2007：36-39页
[31]田广.基于视觉的行人检测和跟踪技术的研究.上海交通大学博士学位论文,2007：53-59页
[32]孙昀,刘富强,李志鹏.基于空间梯度直方图的行人检测算法.中国图象图形学报,2008(10)：1825-1828页
[33]常好丽.运动行人检测与跟踪方法研究.西北工业大学硕士学位论文,2006:28-32页
[34]朱谊强.基于Adaboost算法的实时行人检测系统.西北工业大学硕士学位论文,2006：46-50页
[35]王江涛,杨静宇.红外图像中人体实时检测研究.系统仿真学报,2007(19)：4490-4494页
[36]凌云峰,朱齐丹,吴自新.全景视觉图像柱面理论展开算法实现及其改进.应用科技,2006(9)：4-6页
[37]张帆,朱齐丹,徐光.基于前向映射的全景视觉图像解算方法.光电工程,2009(1):19-25页
[38]原新,王亮,朱齐丹.基于全向视觉传感器的图像解算方法研究.哈尔滨工业大学学报,2006(12)：2158-2161页
[39]Xiangjian He, Jianmin Li, Yan Chen. Local Binary Patterns for Human Detection on Hexagonal Structure.9th IEEE International Symposium on Multimedia,2007:65-71P
[40]Yadong Mu, Shuicheng Yan, Yi Liu. Discriminative local binary patterns for human detection in personal album. Computer Vision and Pattern Recognition,.2008: 121-128P
[41]Sabzmeydani P, Mori G. Detecting Pedestrians by Learning Shapelet Features. Computer Vision and Pattern Recognition,2007:110-118P
[42]Dalal N. Finding people in images and videos. French National Institute for Research in Computer Science and Control,2006:56-63P
[43]Y. Freund and R. Schapire, A Decision-Theoretic Generalization of On-Line Learning and an Application to Boosting. Proceedings of the Second European Conference on Computational Learning Theory 1995:106-110P
[44]V.Vapnik. The Nature of Statistical Learning Theory. NewYork:Spring Press, 1995:125-129P
[45]边肇祺,张学工.模式识别.北京：清华大学出版社,2000:292页
[46]刘广利.基于支持向量机的经济预警方法研究.中国农业大学博士学位论文,2003：23-26页
[47]P.Viola and M.Jones Rapid object detection using a boosted cascade of simple features..Institute of Electrical and Eleetronics Engineers Computer Society, 2001:511-518P
[48]周柯.基于HOG特征的图像人体检测技术的研究与实现.华中科技大学硕士学位论文,2008:60-65页
[49]Comaniciu D. An algorithm for data-driven bandwidth selection. Pattern Analysis and Machine Intelligence,2003,5(2):281-288P
[50]Comaniciu D. Nonparametric information fusion for motion estimation. Computer Vision and Pattern Recognition.2003:59-66P
[51]张舒,褚艳丽.GPU高性能运算之CUDA.北京：中国水利水电出版社,2009
[52]席志红,张志,王帆.全方位视觉大图像并行解算方法研究.应用科技,2009(02):9-12页
[53]Bilgic B, Horn B K P, Masaki I. Efficient integral image computation on the GPU. Intelligent Vehicles Symposium.2010:528-533P
[54]Bilgic B, Horn B K P, Masaki I. Fast human detection with cascaded ensembles on the GPU. Intelligent Vehicles Symposium.2010:325-332P
[55]白洪涛.基于GPU的高性能并行算法研究.吉林大学博士学位论文,2010:79-86页
[56]邹岩,杨志义,张凯龙.CUDA并行程序的内存访问优化技术研究.计算机测量与控制.2009(12)：2504-2506页.
[2]Paul E. Rybski, Fernando de la Torre, Raju Patil, Carlos Vallespi, Manuela M. Veloso, Brett Browning. CAMEO:Camera Assisted Meeting Event Observer. Robotics and Automation. Proceedings of IEEE ICRA,2004:1634-1639P
[3]Patil R, Rybski P E, Kanade T. People detection and tracking in high resolution panoramic video mosaic. Intelligent Robots and Systems, Proceedings.of IEEE/RSJ. 2004:1323-1328P
[4]Chakravarty P, Jarvis R. Panoramic Vision and Laser Range Finder Fusion for Multiple Person Tracking. Intelligent Robots and Systems, Proceedings of IEEE/RSJ. 2006:2949-2954P
[5]Koyasu H, Miura J, Shirai Y. Real-time omnidirectional stereo for obstacle detection and tracking in dynamic environments. Intelligent Robots and Systems, Proceedings of IEEE/RSJ.2001:31-36P
[6]Baba A, Chatila R. Dynamic Targets Detection for Robotic Applications Using Panoramic Vision System. Recent Progress in Robotics:Viable Robotic Service to Human.2008:215-227P
[7]曾吉勇苏显渝.双曲面折反射全景成像系统.光学学报,2003(9):1138-1142页
[8]曾吉勇苏显渝.折反射全景成像系统.激光杂志,2004(6):62-64页
[9]魏芳,董再励,孙茂相,王晓蕾.用于移动机器人的视觉全局定位系统研究.机器人,2001(5)：400-403页
[10]张尧陈卫东.一个基于全景视觉的移动机器人导航系统的设计与实现.机器人,2005(2)：173-177页
[11]王景川陈卫东,曹其新.基于全景视觉与里程计的移动机器人自定位方法研究.机器人,2005(1)：41-45页
[12]白剑,牛爽,杨国光.全景光学环带凝视成像技术.红外与激光工程,2006(3):331-335页
[13]肖潇,杨国光.全景成像技术的现状和进展.光学仪器,2007(4)：84-89页
[14]肖潇,杨国光,白剑.基于球面透视投影约束的全景环形透镜畸变校正.光学学报,2008(4)：675-680页
[15]席志红.全方位视觉技术及其在智能移动机器人等领域的应用研究.哈尔滨工程大学博士学位论文,2006:12-16页
[16]夏桂华,王博,朱齐丹.改进SIFT用于全景视觉移动机器人定位.计算机工程与应用,2010(18)：196-198页.
[17]Viola P, Jones M J, Snow D. Detecting pedestrians using patterns of motion and appearance. Computer Vision,2003:734-741P
[18]N.Dalal and B.Triggs Histograms of oriented gradients for human detection. SanDiego, CA, UnitedStates,2005.Institute of Electrical and Eleetronics Engineers Computer Soeiety, Piscataway, NJ 08855-1331, UnitedStates,886-893P
[19]Qiang Zhu, Mei-Chen Yeh, Kwang-Ting Cheng. Fast Human Detection Using a Cascade of Histograms of Oriented Gradients. Computer Vision and Pattern Recognition,2006:1491-1498P
[20]He X, Li J, Chen Y. Local Binary Patterns with Mahalanobis Distance Maps for Human Detection. Image and Signal Processing.2008:520-524P
[21]Shet V D, Neumann J, Ramesh V. Bilattice-based Logical Reasoning for Human Detection. Computer Vision and Pattern Recognition.2007:1-8P
[22]Ying Wu, Ting Yu. A field model for human detection and tracking. Pattern Analysis and Machine Intelligence,2006,28(5):753-765P
[23]Ying Wu, Ting Yu, Gang Hua. A statistical field model for pedestrian detection. Computer Vision and Pattern Recognition,2005:1023-1030P
[24]Zeng C, Ma H, Ming A. Fast human detection using mi-sVM and a cascade of HOG-LBP features. Image Processing (ICIP),2010:3845-3848P
[25]李同治,丁晓青,王生进.利用级联SVM的人体检测方法.中国图象图形学报,2008(3)：566-570页
[26]贾慧星,章毓晋.车辆辅助驾驶系统中基于计算机视觉的行人检测研究综述[J.自动化学报,2007(1)：84-90页
[27]胡斌,王生进,丁晓青.基于部位检测和子结构组合的行人检测方法.计算机科学,2009(11)：242-246页
[28]许言午.面向行人检测的组合分类计算模型与应用研究.中国科学技术大学博士学位论文,2009:26-30页
[29]杨志辉.基于多尺度方向特征的行人检测算法.哈尔滨工程大学硕士学位论文,2009：36-41页
[30]吴贻军.基于视频序列的运动人体检测算法研究.浙江大学硕士学位论文,2007：36-39页
[31]田广.基于视觉的行人检测和跟踪技术的研究.上海交通大学博士学位论文,2007：53-59页
[32]孙昀,刘富强,李志鹏.基于空间梯度直方图的行人检测算法.中国图象图形学报,2008(10)：1825-1828页
[33]常好丽.运动行人检测与跟踪方法研究.西北工业大学硕士学位论文,2006:28-32页
[34]朱谊强.基于Adaboost算法的实时行人检测系统.西北工业大学硕士学位论文,2006：46-50页
[35]王江涛,杨静宇.红外图像中人体实时检测研究.系统仿真学报,2007(19)：4490-4494页
[36]凌云峰,朱齐丹,吴自新.全景视觉图像柱面理论展开算法实现及其改进.应用科技,2006(9)：4-6页
[37]张帆,朱齐丹,徐光.基于前向映射的全景视觉图像解算方法.光电工程,2009(1):19-25页
[38]原新,王亮,朱齐丹.基于全向视觉传感器的图像解算方法研究.哈尔滨工业大学学报,2006(12)：2158-2161页
[39]Xiangjian He, Jianmin Li, Yan Chen. Local Binary Patterns for Human Detection on Hexagonal Structure.9th IEEE International Symposium on Multimedia,2007:65-71P
[40]Yadong Mu, Shuicheng Yan, Yi Liu. Discriminative local binary patterns for human detection in personal album. Computer Vision and Pattern Recognition,.2008: 121-128P
[41]Sabzmeydani P, Mori G. Detecting Pedestrians by Learning Shapelet Features. Computer Vision and Pattern Recognition,2007:110-118P
[42]Dalal N. Finding people in images and videos. French National Institute for Research in Computer Science and Control,2006:56-63P
[43]Y. Freund and R. Schapire, A Decision-Theoretic Generalization of On-Line Learning and an Application to Boosting. Proceedings of the Second European Conference on Computational Learning Theory 1995:106-110P
[44]V.Vapnik. The Nature of Statistical Learning Theory. NewYork:Spring Press, 1995:125-129P
[45]边肇祺,张学工.模式识别.北京：清华大学出版社,2000:292页
[46]刘广利.基于支持向量机的经济预警方法研究.中国农业大学博士学位论文,2003：23-26页
[47]P.Viola and M.Jones Rapid object detection using a boosted cascade of simple features..Institute of Electrical and Eleetronics Engineers Computer Society, 2001:511-518P
[48]周柯.基于HOG特征的图像人体检测技术的研究与实现.华中科技大学硕士学位论文,2008:60-65页
[49]Comaniciu D. An algorithm for data-driven bandwidth selection. Pattern Analysis and Machine Intelligence,2003,5(2):281-288P
[50]Comaniciu D. Nonparametric information fusion for motion estimation. Computer Vision and Pattern Recognition.2003:59-66P
[51]张舒,褚艳丽.GPU高性能运算之CUDA.北京：中国水利水电出版社,2009
[52]席志红,张志,王帆.全方位视觉大图像并行解算方法研究.应用科技,2009(02):9-12页
[53]Bilgic B, Horn B K P, Masaki I. Efficient integral image computation on the GPU. Intelligent Vehicles Symposium.2010:528-533P
[54]Bilgic B, Horn B K P, Masaki I. Fast human detection with cascaded ensembles on the GPU. Intelligent Vehicles Symposium.2010:325-332P
[55]白洪涛.基于GPU的高性能并行算法研究.吉林大学博士学位论文,2010:79-86页
[56]邹岩,杨志义,张凯龙.CUDA并行程序的内存访问优化技术研究.计算机测量与控制.2009(12)：2504-2506页.