基于非采样高斯差分金字塔的多尺度融合边缘检测
|
摘要
针对传统边缘检测算法抗噪性差、边缘连续度低、细节边缘冗余,对运动目标检测应用领域的适用性差等缺点,基于图像多尺度的思想,结合小尺度图像边缘信息准确,大尺度图像抗噪性强、边缘冗余度低的优点,提出一种基于非采样高斯差分金字塔的多尺度融合边缘检测算法。算法首先对图像进行非采样高斯金字塔分解得到多尺度图像,同时在分解过程实现基于高斯差分算子的边缘检测,得到多尺度边缘图像。最后采用多尺度图像边缘融合策略实现多尺度边缘融合。通过实验对算法的有效性进行验证:通过对边缘融合结果进行Abdou-Pratt品质因数分析,表明该算法抗噪性强,边缘定位准确;连续度分析结果表明该算法在降低边缘冗余度的同时保留了主要边缘,且边缘连续度较高;车辆检测实验结果表明基于该算法得到的车辆检测结果准确度较高。
Traditional edge detection algorithms are low in noise immunity,high in edge continuity,redundant in edge details,and poor in the applicability for moving target detection. In order to treat these problems,a multiscale fused edge detection algorithm was proposed based on non-sampling Difference of Gaussian( Do G) Pyramid. With the concept of multiscale image,the algorithm combined the accurate edge information in small scale image and strong noise immunity and low edge redundancy in large scale image. Experiments verified the effectiveness of the proposed algorithm. Analysis of Abdou-Pratt quality factor on the fused edge image showed that the proposed algorithm is strong in noise immunity and accurate in edge location. Analysis of edge continuity showed that the proposed algorithm decreases edge continuity while keeps main edges,and obtains edges with higher continuity. Experiment result of vehicle detection showed that the vehicle detection result based on the proposed algorithm has higher accuracy.
Traditional edge detection algorithms are low in noise immunity,high in edge continuity,redundant in edge details,and poor in the applicability for moving target detection. In order to treat these problems,a multiscale fused edge detection algorithm was proposed based on non-sampling Difference of Gaussian( Do G) Pyramid. With the concept of multiscale image,the algorithm combined the accurate edge information in small scale image and strong noise immunity and low edge redundancy in large scale image. Experiments verified the effectiveness of the proposed algorithm. Analysis of Abdou-Pratt quality factor on the fused edge image showed that the proposed algorithm is strong in noise immunity and accurate in edge location. Analysis of edge continuity showed that the proposed algorithm decreases edge continuity while keeps main edges,and obtains edges with higher continuity. Experiment result of vehicle detection showed that the vehicle detection result based on the proposed algorithm has higher accuracy.
引文
[1]Saket Bhardwaj,Ajay Mittal.A survey on various edge detector techniques[C]//Proceedings of the 2nd International Conference on Computer,Communication,Control and Information Technology.Netherland:Elsevier Ltd,2012(4):220-226.
[2]Jiang Xing.The research on key problems of human motion target detection and tracking[D]Changsha:Central South University of Forestry and Technology,2013.[蒋钘.人体运动目标检测与跟踪关键问题研究[D].长沙:中南林业科技大学,2013.]
[3]Bin Fenglin,Yi Ming Chan,Li Chenfu,et al.Integrating appearance and edge features for sedan vehicle detection in the blind-Spot area[J].IEEE Transactions on Intelligent Transportation Systems,2012,13(2):737-747.
[4]Soo Siang Teoh,Thomas Brunl.Symmetry-based monocular vehicle detection system[J].Machine Vision and Applications,2012,23(5):831-842.
[5]Aytekin B,Altug E.Increasing driving safety with a multiple vehicle detection and tracking system using ongoing vehicle shadow information[C]//Proceedings of the IEEE International Conference on Systems Man and Cybernetics.Istanbul,Turkey:IEEE,2010:3650-3656.
[6]Li Zhang.Color edge detection by using the center line extraction method[J].Journal of Computers,2014,9(4):922-928.
[7]Luo Y,Duraiswami R.Canny edge detection on NVIDIA CUDA[C]//Proceedings of the IEEE Computer Vision and Pattern Recognition Workshops.Anchorage,US:IEEE,2008:1-8.
[8]Hang Zhou,Dan Peng,Xin Wang,et al.A novel algorithm of edge location based on omni-directional and multi-scale MM[J].Journal of Computers,2014,9(4):990-997.
[9]Tao Wang,Na Wei.Multi-scale mathematical morphology based image edge detection[C]//Proceedings of the Second International Conference on Intelligent System Design&Engineering Application.Sanya,China:IEEE,2012(674):1060-1062.
[10]Qian Xu,Srenivas Varadarajan,Chaitali Chakrabarti.A distributed Canny edge detector:Algorithm and FPGA implementation[J].IEEE Transactions on Image Processing,2014,23(7):2944-2960.
[11]Lowe D G.Object recognition from local scale invariant features[C]//Proceedings of the International Conference on Computer Vision.Kerkyra,Greece:IEEE,1999(2):1150-1157.
[12]Lourenco L H A.Efficient implementation of canny edge detection filter for ITK using CUDA[C]//Proceedings of the 13th Symp.Comput.Syst.Petropolis,Brazil:IEEE,2012:33-40.
[13]Zhang Qi.Image edge detection algorithm based on multiscales wavelet transform[D].Changchun:Jilin University,2011.[张琦.基于多尺度小波变换的图像边缘检测算法[D].长春:吉林大学,2011.]
[14]Charaa S,Ellouze N.Multiscale product edge detection in different colour spaces[C]//Proceedings of the 11th International Conference on Information Technology:New Generations.Washington,USA:IEEE Computer Society,2014:660-664.
[15]Jianxin Zhang,Wei Chang.Edge detection based on general gray correlation and Lo G operator[C]//Proceedings of the International Conference on Artificial Intelligence and Computational Intelligence.Sanya,China:IEEE,2010(106):480-483.
[16]Li Xue.Performance evaluation in edge detection algorithms of gray images[D].Shenyang:Shenyang University of Technology,2010.[李雪.灰度图像边缘检测算法的性能评价[D].沈阳:沈阳工业大学,2010.]
[17]Williams I,Bowring N,Svoboda D.A performance evaluation of statistical tests for edge detection in textured images[J].Computer Vision and Image Understanding,2014,122:115-30.
[18]Abdou I E,Pratt W K.Quantitative design and evaluation of enhancement/thresholding edge detectors[C]//Proceedings of the IEEE.Washington,USA:IEEE,1979,67(5):753-763.
[19]Mo Shaoqing.Research on edge detection and its evaluation[D].Tianjin:Tianjin University,2011.[磨少清.边缘检测及其评价方法的研究[D].天津:天津大学,2011.]
[20]Standard Test Image Database[2014-3-16].http://www.image processing place.com/.
[2]Jiang Xing.The research on key problems of human motion target detection and tracking[D]Changsha:Central South University of Forestry and Technology,2013.[蒋钘.人体运动目标检测与跟踪关键问题研究[D].长沙:中南林业科技大学,2013.]
[3]Bin Fenglin,Yi Ming Chan,Li Chenfu,et al.Integrating appearance and edge features for sedan vehicle detection in the blind-Spot area[J].IEEE Transactions on Intelligent Transportation Systems,2012,13(2):737-747.
[4]Soo Siang Teoh,Thomas Brunl.Symmetry-based monocular vehicle detection system[J].Machine Vision and Applications,2012,23(5):831-842.
[5]Aytekin B,Altug E.Increasing driving safety with a multiple vehicle detection and tracking system using ongoing vehicle shadow information[C]//Proceedings of the IEEE International Conference on Systems Man and Cybernetics.Istanbul,Turkey:IEEE,2010:3650-3656.
[6]Li Zhang.Color edge detection by using the center line extraction method[J].Journal of Computers,2014,9(4):922-928.
[7]Luo Y,Duraiswami R.Canny edge detection on NVIDIA CUDA[C]//Proceedings of the IEEE Computer Vision and Pattern Recognition Workshops.Anchorage,US:IEEE,2008:1-8.
[8]Hang Zhou,Dan Peng,Xin Wang,et al.A novel algorithm of edge location based on omni-directional and multi-scale MM[J].Journal of Computers,2014,9(4):990-997.
[9]Tao Wang,Na Wei.Multi-scale mathematical morphology based image edge detection[C]//Proceedings of the Second International Conference on Intelligent System Design&Engineering Application.Sanya,China:IEEE,2012(674):1060-1062.
[10]Qian Xu,Srenivas Varadarajan,Chaitali Chakrabarti.A distributed Canny edge detector:Algorithm and FPGA implementation[J].IEEE Transactions on Image Processing,2014,23(7):2944-2960.
[11]Lowe D G.Object recognition from local scale invariant features[C]//Proceedings of the International Conference on Computer Vision.Kerkyra,Greece:IEEE,1999(2):1150-1157.
[12]Lourenco L H A.Efficient implementation of canny edge detection filter for ITK using CUDA[C]//Proceedings of the 13th Symp.Comput.Syst.Petropolis,Brazil:IEEE,2012:33-40.
[13]Zhang Qi.Image edge detection algorithm based on multiscales wavelet transform[D].Changchun:Jilin University,2011.[张琦.基于多尺度小波变换的图像边缘检测算法[D].长春:吉林大学,2011.]
[14]Charaa S,Ellouze N.Multiscale product edge detection in different colour spaces[C]//Proceedings of the 11th International Conference on Information Technology:New Generations.Washington,USA:IEEE Computer Society,2014:660-664.
[15]Jianxin Zhang,Wei Chang.Edge detection based on general gray correlation and Lo G operator[C]//Proceedings of the International Conference on Artificial Intelligence and Computational Intelligence.Sanya,China:IEEE,2010(106):480-483.
[16]Li Xue.Performance evaluation in edge detection algorithms of gray images[D].Shenyang:Shenyang University of Technology,2010.[李雪.灰度图像边缘检测算法的性能评价[D].沈阳:沈阳工业大学,2010.]
[17]Williams I,Bowring N,Svoboda D.A performance evaluation of statistical tests for edge detection in textured images[J].Computer Vision and Image Understanding,2014,122:115-30.
[18]Abdou I E,Pratt W K.Quantitative design and evaluation of enhancement/thresholding edge detectors[C]//Proceedings of the IEEE.Washington,USA:IEEE,1979,67(5):753-763.
[19]Mo Shaoqing.Research on edge detection and its evaluation[D].Tianjin:Tianjin University,2011.[磨少清.边缘检测及其评价方法的研究[D].天津:天津大学,2011.]
[20]Standard Test Image Database[2014-3-16].http://www.image processing place.com/.