基于前景分割的行人检测方法
|
摘要
针对背景复杂情况下行人检测误检率较大的问题,提出一种新的基于前景分割的行人检测方法.本方法在样本训练过程中,通过对图像的初始轮廓线进行有向分水岭转换,然后由超度量轮廓图算法得到图像内一个个封闭的区域,把得到的封闭区域与设定框进行比较,区分封闭区域属于前景还是背景,进而把前景目标分割出来并进行训练;测试时,把待检测图像中的检测区域进行前景分割,求出前景的HOG特征并用SVM分类,确定检测区域内是否有行人.这样保证了在训练阶段和检测阶段都去除了背景噪声的影响,实验结果表明,提出的方法能有效的提高检测精度.
The complex background will greatly affect the test accuracy of human detection. In order to improve the accuracy of human detection,in this paper a new method of Foreground Segmentation has been proposed. This method is divided into tw o phases, in the sample training phase,through Oriented Watershed Transform and Ultrametric Contour Map,many closed regions in the image can be got,then w e compare these closed regions w ith a box w hich has been set,and determine these closed regions is foreground or not. The foreground in the image can be got and trained. During the testing phase,the area in the test image w hich need to be detected can be segmentalized and the foreground can be got,then w e can get the HOG of the foreground. By SVM,w e know that there is a human in the area or not. So the foreground characteristic can be calculated w hich have no background noise in the sample training phase and testing phase. The experimental results show that this approach is effective in improving detection accuracy.
The complex background will greatly affect the test accuracy of human detection. In order to improve the accuracy of human detection,in this paper a new method of Foreground Segmentation has been proposed. This method is divided into tw o phases, in the sample training phase,through Oriented Watershed Transform and Ultrametric Contour Map,many closed regions in the image can be got,then w e compare these closed regions w ith a box w hich has been set,and determine these closed regions is foreground or not. The foreground in the image can be got and trained. During the testing phase,the area in the test image w hich need to be detected can be segmentalized and the foreground can be got,then w e can get the HOG of the foreground. By SVM,w e know that there is a human in the area or not. So the foreground characteristic can be calculated w hich have no background noise in the sample training phase and testing phase. The experimental results show that this approach is effective in improving detection accuracy.
引文
[1]Gong Yong-gang.License plate recognition algorithm with sparse representation based on local HOG[J].Computer Simulation,2011,28(4):367-369.
[2]Cheng Guang-tao,Chen Xue,Guo Zhao-zhuang.Pedestrian detection method of vision based on HOG features[J].Transducer and Microsystem Technologies,2011,30(7):68-70.
[3]Gao En-yang,Liu Wei-jun,Wang Tian-ran.Interactive mesh segmentation based on graph laplacian[C].Applied Mechanics and Materials.Switzerland:Technology for Manufacturing Systems,2011:1535-1540.
[4]Zhou Jin-zhi,Wang Juan.Improvement of human detection method based on HOG[J].Software Guide,2011,10(4):76-78.
[5]Dalai N,Triggs B.Histograms of oriented gradients for human detection[C].Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition,San Diego:United states,2005:886-893.
[6]Yao Xue-qin,Li Xiao-hua.Pedestrian detection based on edge symmetry and HOG[J].Computer Engineering,2011,37(20):1-4.
[7]Niu Jie,Qian Kun.Pedestrian detection based on multiscale and multishape HOG features[J].Computer Technology and Development,2011,21(9):99-102.
[8]Li Zhen,Wei Zhi-qiang.Pedestrian detection based on adaptive background modeling[J].Journal of System Simulation,2009,21(1):61-64.
[9]Fowlkes C,Martin D and Malik J.Learning affinity functions for image segmentation:combining patch-based and gradient-based approaches[C].Computer Vision and Pattern Recognition,Madison:IEEE,2003:54-61.
[10]Arbelaze P,Maire M,Fowlkes C.Contour detection and hierarchical image segmentation[J].Pattern Analysis and Machine Intelligence,2011,33(5):898-916.
[11]Lim J J,Arbelaze P,Malik J.Recognition using regions[C].Computer Vision and Pattern Recognition,Athens:IEEE,2009:1030-1037.
[12]Ren Xu,Gu Cheng-cheng.Hand gesture recognition based on HOG characters and SVM[J].Bulletin of Science and Technology,2011,27(2):211-214.
[1]龚永罡.基于局部HOG特征的稀疏表达车牌识别算法[J].计算机仿真,2011,28(4):367-369.
[2]程广涛,陈雪,郭照庄.基于HOG特征的行人视觉检测方法[J].传感器与微系统,2011,30(7):68-70.
[4]周金芝,王娟.基于HOG的人体检测方法的改进[J].软件导刊,2011,10(4):76-78.
[6]姚雪琴,李晓华.基于边缘对称性和HOG的行人检测[J].计算机工程,2011,37(20):1-4.
[7]牛杰,钱堃.基于多尺度-多形状HOG特征的行人检测方法[J].计算机技术与发展,2011,21(9):99-102.
[8]李臻,魏志强.基于自适应背景模型的行人检测方法[J].系统仿真学报,2009,21(1):61-64.
[12]任彧,顾成成.基于HOG特征和SVM的手势识别[J].科技通报,2011,27(2):211-214.
[2]Cheng Guang-tao,Chen Xue,Guo Zhao-zhuang.Pedestrian detection method of vision based on HOG features[J].Transducer and Microsystem Technologies,2011,30(7):68-70.
[3]Gao En-yang,Liu Wei-jun,Wang Tian-ran.Interactive mesh segmentation based on graph laplacian[C].Applied Mechanics and Materials.Switzerland:Technology for Manufacturing Systems,2011:1535-1540.
[4]Zhou Jin-zhi,Wang Juan.Improvement of human detection method based on HOG[J].Software Guide,2011,10(4):76-78.
[5]Dalai N,Triggs B.Histograms of oriented gradients for human detection[C].Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition,San Diego:United states,2005:886-893.
[6]Yao Xue-qin,Li Xiao-hua.Pedestrian detection based on edge symmetry and HOG[J].Computer Engineering,2011,37(20):1-4.
[7]Niu Jie,Qian Kun.Pedestrian detection based on multiscale and multishape HOG features[J].Computer Technology and Development,2011,21(9):99-102.
[8]Li Zhen,Wei Zhi-qiang.Pedestrian detection based on adaptive background modeling[J].Journal of System Simulation,2009,21(1):61-64.
[9]Fowlkes C,Martin D and Malik J.Learning affinity functions for image segmentation:combining patch-based and gradient-based approaches[C].Computer Vision and Pattern Recognition,Madison:IEEE,2003:54-61.
[10]Arbelaze P,Maire M,Fowlkes C.Contour detection and hierarchical image segmentation[J].Pattern Analysis and Machine Intelligence,2011,33(5):898-916.
[11]Lim J J,Arbelaze P,Malik J.Recognition using regions[C].Computer Vision and Pattern Recognition,Athens:IEEE,2009:1030-1037.
[12]Ren Xu,Gu Cheng-cheng.Hand gesture recognition based on HOG characters and SVM[J].Bulletin of Science and Technology,2011,27(2):211-214.
[1]龚永罡.基于局部HOG特征的稀疏表达车牌识别算法[J].计算机仿真,2011,28(4):367-369.
[2]程广涛,陈雪,郭照庄.基于HOG特征的行人视觉检测方法[J].传感器与微系统,2011,30(7):68-70.
[4]周金芝,王娟.基于HOG的人体检测方法的改进[J].软件导刊,2011,10(4):76-78.
[6]姚雪琴,李晓华.基于边缘对称性和HOG的行人检测[J].计算机工程,2011,37(20):1-4.
[7]牛杰,钱堃.基于多尺度-多形状HOG特征的行人检测方法[J].计算机技术与发展,2011,21(9):99-102.
[8]李臻,魏志强.基于自适应背景模型的行人检测方法[J].系统仿真学报,2009,21(1):61-64.
[12]任彧,顾成成.基于HOG特征和SVM的手势识别[J].科技通报,2011,27(2):211-214.