点线特征融合的误匹配剔除算法
|
摘要
特征匹配作为计算机视觉的一项关键技术而备受关注。近年来,基于描述子的特征点匹配技术取得了一系列突破性进展,但曲线长度不一、端点定位不准确以及周围包含的重复性纹理较多等因素,导致了曲线匹配研究依旧是一个极具挑战性的热点研究课题,且现有曲线匹配方法大多出现匹配总数少、匹配正确率低的问题。为增加特征匹配的总数和正确率,利用特征点和特征曲线的位置关系提出一种点线特征融合的误匹配剔除算法(Point Line feature Fusion,PLF)。首先定义点到曲线的距离,利用点、曲线描述子提取图像的点、线特征;其次确定落入匹配曲线对应支撑区域内的匹配点对,并根据匹配点组和曲线间的距离约束剔除错误曲线匹配;最后利用点线距离约束剔除匹配曲线支撑区域内的错误点匹配。实验选取了3种不同的点线组合,即SIFT技术提取的点特征分别与IOCD曲线描述子、IOMSD曲线描述子、GOCD曲线描述子提取的曲线特征相融合,验证算法对多种点、线描述子具有适用性,且该算法不仅适用于特征点与特征曲线的融合,亦适用于特征点与特征直线的融合,从而验证了其对多种图像特征具有适用性。实验结果表明,在旋转、视角变化、光照变化、压缩、噪音、模糊等变换条件下,该算法均能有效提高曲线特征匹配的匹配总数和匹配正确率,同时提高点匹配的正确率。
Image feature matching plays an important role in computer vision,the feature point matching technology based on descriptors have made a series of achievements,since curves have different lengths,incorrect position of endpoints and contain lots of relative texture around neighbor,the research of feature curve matching is still a challenging topic,and many curve matching methods have the problem of fewer matches and low accuracy of feature matching.To improve the total number and accuracy of feature matching,this paper proposed a novel Point Line feature Fusion(PLF)algorithm based on the location relationship between feature points and feature curves.Firstly,it defines the distance from a point to a curve,and obtains the matched points and curves using point and curve descriptors respectively from the images.Secondly,it determines the matched point pairs in the support areas of one pair of matched curves,and eliminate the mismatch of curves according to the distance constraints between the obtained matched points and the curve.Then,it removes the mismatch of points according to the distance constraint between the point and the curve.Three combinations of points and curves have been used in the experiment,which are the points extracted by SIFT and the curves extracted by IOCD curve descriptor,the points extracted by SIFT and the curves extracted by IOMSD curve descriptor,the points extracted by SIFT and the curves extracted by GOCD curve descriptor.The method has applicability to many kinds of point and curve descriptors,it is not only suitable for the points and curves,but also for points and lines,it has applicability to many kinds of features.Experimental results show that the proposed algorithm can effectively improve the total number and accuracy of feature matching,and also increase the accuracy of point matching under image rotation,viewpoint change,illumination change,JPEG compression,noise and image blur.
Image feature matching plays an important role in computer vision,the feature point matching technology based on descriptors have made a series of achievements,since curves have different lengths,incorrect position of endpoints and contain lots of relative texture around neighbor,the research of feature curve matching is still a challenging topic,and many curve matching methods have the problem of fewer matches and low accuracy of feature matching.To improve the total number and accuracy of feature matching,this paper proposed a novel Point Line feature Fusion(PLF)algorithm based on the location relationship between feature points and feature curves.Firstly,it defines the distance from a point to a curve,and obtains the matched points and curves using point and curve descriptors respectively from the images.Secondly,it determines the matched point pairs in the support areas of one pair of matched curves,and eliminate the mismatch of curves according to the distance constraints between the obtained matched points and the curve.Then,it removes the mismatch of points according to the distance constraint between the point and the curve.Three combinations of points and curves have been used in the experiment,which are the points extracted by SIFT and the curves extracted by IOCD curve descriptor,the points extracted by SIFT and the curves extracted by IOMSD curve descriptor,the points extracted by SIFT and the curves extracted by GOCD curve descriptor.The method has applicability to many kinds of point and curve descriptors,it is not only suitable for the points and curves,but also for points and lines,it has applicability to many kinds of features.Experimental results show that the proposed algorithm can effectively improve the total number and accuracy of feature matching,and also increase the accuracy of point matching under image rotation,viewpoint change,illumination change,JPEG compression,noise and image blur.
引文
[1] PAN Z,LI Z,FAN H,et al.Feature Based Local Binary Pattern for Rotation Invariant Texture Classification[J].Expert Systems with Applications,2017,88:238-248.
[2] ZHANG S,TIAN Q,LU K,et al.Edge-SIFT:Discriminative Binary Descriptor for Scalable Partial-Duplicate Mobile Search[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,2013,22(7):2889-2902.
[3] ZENG L,ZHANG S,ZHANG J,et al.Dynamic image mosaic via SIFT and dynamic programming[J].Machine Vision&Applications,2014,25(5):1271-1282.
[4] LIU Y,HE F,ZHU X,et al.The Improved Characteristics of Bionic Gabor Representations by Combining with SIFT Keypoints for Iris Recognition[J].Journal of Bionic Engineering,2015,12(3):504-517.
[5] LI J,ALLINSON N.Building Recognition Using Local Oriented Features[J].IEEE Transactions on Industrial Informatics,2013,9(3):1697-1704.
[6] LI Z,GAO S,KE N.Robust Object Tracking Based on Adaptive Templates Matching via the Fusion of Multiple Features[J].Journal of Visual Communication&Image Representation,2017,44:1-20.
[7] LOWE D G.Distinctive Image Features from Scale-Invariant Keypoints[J].International Journal of Computer Vision,2004,60(2):91-110.
[8] KE Y,SUKTHANKAR R.PCA-SIFT:a more distinctive representation for local image descriptors[C]∥IEEE Computer Society Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2004:506-513.
[9] BAY H,TUYTELAARS T,GOOL L V.SURF:Speeded Up Robust Features[C]∥European Conference on Computer Vision.Springer,Berlin,Heidelberg,2006:404-417.
[10]CALONDER M,LEPETIT V,OZUYSAL M,et al.BRIEF:Computing a Local Binary Descriptor very Fast[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2012,34(7):1281.
[11]RUBLEE E,RABAUD V,KONOLIGE K,et al.ORB:An efficient alternative to SIFT or SURF[C]∥IEEE International Conference on Computer Vision.IEEE,2012:2564-2571.
[12]LEUTENEGGER S,CHLI M,SIEGWART R Y.BRISK:Binary Robust invariant scalable keypoints[C]∥IEEE International Conference on Computer Vision.IEEE,2012:2548-2555.
[13]WANG Z,WU F,HU Z.MSLD:A robust descriptor for line matching[J].Pattern Recognition,2009,42(5):941-953.
[14]FAN B,WU F,HU Z.Aggregating gradient distributions into intensity orders:A novel local image descriptor[C]∥IEEE Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2011:2377-2384.
[15]WANG Z H,ZHI S S,LIU H M.IOMSD:Intensity order based mean-standed deviation descriptor[J].Pattern Recognition and Artificial Intelligence,2012,26(4):409-416.(in Chinese)王志衡,智珊珊,刘红敏.基于亮度序的均值标准差描述子[J].模式识别与人工智能,2012,26(4):409-416.
[16]LIU H M,ZHI S S,WANG Z H.IOCD:Intensity Order Curve Descriptor[J].International Journal of Pattern Recognition and Artificial Intellgence,2013,27(7):1355011.
[17]LOURAKIS M I A,HALKIDIS S T,ORPHANOUDAKIS S C.Matching disparate views of planar surfaces using projective invariants[J].Image&Vision Computing,2000,18(9):673-683.
[18]FAN B,WU F,HU Z.Line matching leveraged by point correspondences[C]∥Computer Vision and Pattern Recognition.IEEE,2010:390-397.
[19]ZHANG L,KOCH R.An efficient and robust line segment matching approach based on LBD descriptor and pairwise geometric consistency[J].Journal of Visual Communication&Image Representation,2013,24(7):794-805.
[20]LIU H,CHEN L,WANG Z,et al.GOCD:Gradient Order Curve Descriptor[J].Ieice Transactions on Information&Systems,2017,E100.D(12):2973-2983.
[2] ZHANG S,TIAN Q,LU K,et al.Edge-SIFT:Discriminative Binary Descriptor for Scalable Partial-Duplicate Mobile Search[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,2013,22(7):2889-2902.
[3] ZENG L,ZHANG S,ZHANG J,et al.Dynamic image mosaic via SIFT and dynamic programming[J].Machine Vision&Applications,2014,25(5):1271-1282.
[4] LIU Y,HE F,ZHU X,et al.The Improved Characteristics of Bionic Gabor Representations by Combining with SIFT Keypoints for Iris Recognition[J].Journal of Bionic Engineering,2015,12(3):504-517.
[5] LI J,ALLINSON N.Building Recognition Using Local Oriented Features[J].IEEE Transactions on Industrial Informatics,2013,9(3):1697-1704.
[6] LI Z,GAO S,KE N.Robust Object Tracking Based on Adaptive Templates Matching via the Fusion of Multiple Features[J].Journal of Visual Communication&Image Representation,2017,44:1-20.
[7] LOWE D G.Distinctive Image Features from Scale-Invariant Keypoints[J].International Journal of Computer Vision,2004,60(2):91-110.
[8] KE Y,SUKTHANKAR R.PCA-SIFT:a more distinctive representation for local image descriptors[C]∥IEEE Computer Society Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2004:506-513.
[9] BAY H,TUYTELAARS T,GOOL L V.SURF:Speeded Up Robust Features[C]∥European Conference on Computer Vision.Springer,Berlin,Heidelberg,2006:404-417.
[10]CALONDER M,LEPETIT V,OZUYSAL M,et al.BRIEF:Computing a Local Binary Descriptor very Fast[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2012,34(7):1281.
[11]RUBLEE E,RABAUD V,KONOLIGE K,et al.ORB:An efficient alternative to SIFT or SURF[C]∥IEEE International Conference on Computer Vision.IEEE,2012:2564-2571.
[12]LEUTENEGGER S,CHLI M,SIEGWART R Y.BRISK:Binary Robust invariant scalable keypoints[C]∥IEEE International Conference on Computer Vision.IEEE,2012:2548-2555.
[13]WANG Z,WU F,HU Z.MSLD:A robust descriptor for line matching[J].Pattern Recognition,2009,42(5):941-953.
[14]FAN B,WU F,HU Z.Aggregating gradient distributions into intensity orders:A novel local image descriptor[C]∥IEEE Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2011:2377-2384.
[15]WANG Z H,ZHI S S,LIU H M.IOMSD:Intensity order based mean-standed deviation descriptor[J].Pattern Recognition and Artificial Intelligence,2012,26(4):409-416.(in Chinese)王志衡,智珊珊,刘红敏.基于亮度序的均值标准差描述子[J].模式识别与人工智能,2012,26(4):409-416.
[16]LIU H M,ZHI S S,WANG Z H.IOCD:Intensity Order Curve Descriptor[J].International Journal of Pattern Recognition and Artificial Intellgence,2013,27(7):1355011.
[17]LOURAKIS M I A,HALKIDIS S T,ORPHANOUDAKIS S C.Matching disparate views of planar surfaces using projective invariants[J].Image&Vision Computing,2000,18(9):673-683.
[18]FAN B,WU F,HU Z.Line matching leveraged by point correspondences[C]∥Computer Vision and Pattern Recognition.IEEE,2010:390-397.
[19]ZHANG L,KOCH R.An efficient and robust line segment matching approach based on LBD descriptor and pairwise geometric consistency[J].Journal of Visual Communication&Image Representation,2013,24(7):794-805.
[20]LIU H,CHEN L,WANG Z,et al.GOCD:Gradient Order Curve Descriptor[J].Ieice Transactions on Information&Systems,2017,E100.D(12):2973-2983.