基于仿射不变离散哈希的遥感图像快速目标检测新方法
|
摘要
提出一种基于仿射不变离散哈希的遥感图像快速目标检测新方法.首先使用一种"选择性搜索"的方法生成目标候选框;其次,提出一种基于仿射不变离散哈希(Affine-Invariant Discrete Hashing,AIDH)的目标检测方法,该方法采用具有低存储、高效率优势的监督离散哈希框架,结合仿射不变优化因子,构造仿射不变离散哈希,通过将具有相同语义信息的仿射变换样本约束到相似的二值码空间,实现检测精度的提高;最后采用判别分类器结合非极大值抑制的方法,进一步过滤掉误检目标框,完成目标的精确定位.实验证明,在NWPU VHR-10数据集下,该方法相比于经典目标检测方法和新的哈希方法,在具备高效性的同时,在精度上也得到了保证.
A novel fast object detection method in remote sensing images based on Affine-Invariant Discrete Hashing(AIDH)is proposed.Firstly,the selective search method is introduced for region proposal generation.Secondly,affine-invariant discrete hashing is suggested for object detection.This method uses supervised discrete hashing with the advantage of low storage and high efficiency,jointed with affine-invariant factor,to construct affine-invariant discrete hashing.By constraining the affine transform samples with the same semantic information to the similar binary code space,the method achieves the enhancement on classification precision.Finally,we use discriminating classifier and non-maximum suppression for further filtering false detection of object area and accomplishing accurate object localization.Experiments show that under the dataset of NWPU VHR-10 the proposed method is more efficient than classic detection method and new hash method,and it is also guaranteed in accuracy.
A novel fast object detection method in remote sensing images based on Affine-Invariant Discrete Hashing(AIDH)is proposed.Firstly,the selective search method is introduced for region proposal generation.Secondly,affine-invariant discrete hashing is suggested for object detection.This method uses supervised discrete hashing with the advantage of low storage and high efficiency,jointed with affine-invariant factor,to construct affine-invariant discrete hashing.By constraining the affine transform samples with the same semantic information to the similar binary code space,the method achieves the enhancement on classification precision.Finally,we use discriminating classifier and non-maximum suppression for further filtering false detection of object area and accomplishing accurate object localization.Experiments show that under the dataset of NWPU VHR-10 the proposed method is more efficient than classic detection method and new hash method,and it is also guaranteed in accuracy.
引文
[1]曾接贤,付俊,符祥.特征点和不变矩结合的遥感图像飞机目标识别.中国图象图形学报,2014,19(4):592-602.(Zeng J X,Fu J,Fu X.Aircraft target Recognition in remote sensing images based on distribution of the feature points and invariant moments.Journal of Image and Graphics,2014,19(4):592-602.)
[2]张临杰,张杰,张晰等.任务分配均衡的双参数CFAR舰船检测并行算法.遥感学报,2016,20(2):344-351.(Zhang L J,Zhang J,Zhang X,et al.Task distribution balancing for parallel two-parameter CFAR ship detection.Journal of Remote Sensing,2016,20(2):344-351.)
[3]陈卓,马洪超,张良.改进邻域结构与轮廓分析的LiDAR点云立交桥提取.遥感学报,2013,17(5):1146-1157.(Chen Z,Ma H C,Zhang L.Cloverleaf interchange boundary extraction from airborne LiDAR data based on advanced neighborhood structure and contour analysis.Journal of Remote Sensing,2013,17(5):1146-1157.)
[4] Cheng G,Zhou P C,Yao X W,et al.Object detection in VHR optical remote sensing images via learning rotation-invariant HOG feature∥2016 4th International Workshop on Earth ObservationandRemoteSensingApplications(EORSA).Guangzhou,China:IEEE,2016:433-43.doi:10.1109/EORSA.2016.7552845.
[5] Cheng G,Han J W.A survey on object detection in optical remote sensing images.ISPRS Journal of Photogrammetry and Remote Sensing,2016,117:11-28.doi:10.1016/j.isprsjprs.2016.03.014.
[6] Liu W F,Zhang Z Q,Li S Y,et al.Road detectionbyusinga generalizedHough transform.Remote Sensing,2017,9(6):590.doi:10.3390/rs9060590.
[7] Bulley H N N,Bishop M P,Shroder J F,et al.Integration of classification tree analyses and spatial metrics to assess changes in supraglacial lakes in the Karakoram Himalaya.International Journal of Remote Sensing,2012,34(2):387-411.doi:10.1080/01431161.2012.705915.
[8] Holbling D,Füreder P,Antolini F,et al.A semi-automatedobject-basedapproachfor landslidedetectionvalidatedbypersistent scatterer interferometry measures and landslide inventories.Remote Sensing,2012,4(5):1310-1336.doi:10.3390/rs4051310.
[9] Ok A O.Automated detection of buildings from single VHR multispectral images using shadow information and graph cuts.ISPRS Journal of Photogrammetry and Remote Sensing,2013,86:21-40.doi:10.1016/j.isprsjprs.2013.09.004.
[10]Khadanga G,Jain K,Merugu S.Use of OBIA for extraction of cadastral parcels∥International Conference on Advances in Computing,Communications and Informatics(ICACCI).Jaipur,India:IEEE Press,2016:2226-2230.doi:10.1109/ICACCI.2016.7732382.
[11]Hosang J,Benenson R,Dollr P,et al.What makes for effective detection proposals?IEEE Transactions on Pattern Analysis and Machine Intelligence(TPAMI),2016,38(4):814-830.doi:10.1109/TPAMI.2015.2465908.
[12]Uijlings J R R,VanDeSande K E A,Gevers T,et al.Selective search for object recognition.International Journal of Computer Vision,2013,104(2):154-171.doi:10.1007/s11263-013-0620-5.
[13]Zhao C H,Liu W,Xu Y,et al.A spectral-spatial SVM-based multi-layer learning algorithm for hyperspectralimageclassification. Remote Sensing Letters,2018,9(3):218-227.doi:10.1080/2150704X.2017.1415475.
[14]Wu S L,Chen H D,Bai Y,et al.A remote sensing image classification method based on sparse representation. Multimedia Tools and Applications,2016,75(19):12137-12154.doi:10.1007/s11042-016-3320-7.
[15]Cheng G,Zhou P C,Han J W.Learning rotationinvariant convolutional neural networks for object detection in VHR optical remote sensing images.IEEE Transactions on Geoscience and Remote Sensing,2016,54(12):7405-7415.doi:10.1109/TGRS.2016.2601622.
[16]王小美,张红利.基于测地距离的KNN高光谱遥感图像分类.山西煤炭管理干部学院学报,2013,26(4):135-137.(Wang X M,Zhang H L.Hyperspectral remote sensing image classification using geodesic-based KNN.Journal of Shanxi Coalmining Administrators College,2013,26(4):135-137.)
[17]Wang J D,Zhang T,Song J K,et al.A survey on learning to hash.IEEE Transactions on Pattern Analysis and Machine Intelligence,2018,40(4):769-790.doi:10.1109/TPAMI.2017.2699960.
[18]Weiss Y,Torralba A,Fergus R.Spectral hashing∥International Conference on Neural InformationProcessingSystems(NIPS).Vancouver,Canada:Curran Associates Inc,2008,282(3):1753-1760.
[19]Jiang Q Y,Li W J.Scalable graph hashing with feature transformation∥Proceedings of the 24th International Conference on Artificial Intelligence(ICAI).Buenos Aires,Argentina:AAAI Press.2015:2248-2254.doi:10.1145/1814433.1814442.
[20]Shen F M,Shen C H,Shi Q F,et al.Inductive hashing on manifolds∥2013IEEE Conference on Computer Vision and Pattern Recognition.Portland,OR,USA:IEEE Press,2013:1562-1569.doi:10.1109/CVPR.2013.205.
[21]Liu W,Wang J,Ji R R,et al.Supervised hashing withkernels∥2012 IEEEConference onComputer Vision and Pattern Recognition.Providence,RI,USA:IEEE Press,2012:2074-2081.doi:10.1109/CVPR.2012.6247912.
[22]Norouzi M,Fleet D J.Minimal loss hashing for compact binary codes∥Proceedings of the 28th InternationalConferenceonInternational Conference on Machine Learning.Washington DC,USA:Omni Press,2011:353-360.
[23]Shen F M,Shen C H,Liu W,et al.Supervised discrete hashing∥2015 IEEE Conference on Computer VisionandPatternRecognition(CVPR).Boston,MA,USA:IEEE Press,2015:37-45.doi:10.1109/CVPR.2015.7298598.
[24]沈肖波.多视图嵌入学习方法及其应用研究.博士学位论文.南京:南京理工大学,2017.(Shen X B.Study of multi-view embedding learning techniques with applications.Ph.D.Dissertation.Nanjing:Nanjing University of Science and Technology,2017.)
[25]唐涛.图像仿射不变特征及其在遥感图像目标识别中的应用.硕士学位论文.长沙:国防科学技术大学,2006.(Tang T.Affine invariant feature and its application to target recognition in remote sensing images.Master Dissertation.Changsha:National University of Defense Technology,2006.)
[26]王凤国,冯象初,张小波.平稳小波变换在仿射不变性目标识别中的应用.计算机工程与应用,2007,43(21):239-241.(Wang F G,Feng X C,Zhang X B.Stationary wavelet transform for affineinvariantimageobjectrecognition.Computer Engineering and Applications,2007,43(21):239-241.)
[27]张洁玉,陈强,白小晶等.基于广义典型相关分析的仿射不变特征提取方法.电子与信息学报,2009,31(10):2465-2469.(Zhang J Y,Chen Q,Bai X J,et al.Affine invariant feature extraction algorithmbasedongeneralizedcanonical correlation analysis.Journal of Electronics&Information Technology,2009,31(10):2465-2469.)
[28]高峰,文项坚.利用仿射几何的仿射不变特征提取方法.中国图象图形学报,2011,16(3):389-397.(Gao F,Wen X J.A new method for affine invariants extraction based on affine geometry.Journal of Image and Graphics,2011,16(3):389-397.)
[29]Felzenszwalb P F,Huttenlocher D P.Efficient graph-based image segmentation.International Journal of Computer Vision,2004,59(2):167-181.doi:10.1023/b:visi.0000022288.19776.77.
[30]Long Y,Gong Y P,Xiao Z F,et al.Accurate object localization in remote sensing images based on convolutional neural networks.IEEE Transactions on Geoscience and Remote Sensing,2017,55(5):2486-2498.doi:10.1109/TGRS.2016.2645610.
[31]Felzenszwalb P F,Girshick R B,Mcallester D,et al.Object detection with discriminatively trained part-based models.IEEE Transactions on Pattern Analysis and Machine Intelligence,2010,32(9):1627-1645.doi:10.1109/TPAMI.2009.167.
[32]刘静,郭建,贺遵亮.基于Gist和PHOG特征的场景分类.计算机工程,2015,41(4):232-235,240.(Liu J,Guo J,He Z L.Scene classification based on gist and PHOG feature.Computer Engineering,2015,41(4):232-235,240.)
[2]张临杰,张杰,张晰等.任务分配均衡的双参数CFAR舰船检测并行算法.遥感学报,2016,20(2):344-351.(Zhang L J,Zhang J,Zhang X,et al.Task distribution balancing for parallel two-parameter CFAR ship detection.Journal of Remote Sensing,2016,20(2):344-351.)
[3]陈卓,马洪超,张良.改进邻域结构与轮廓分析的LiDAR点云立交桥提取.遥感学报,2013,17(5):1146-1157.(Chen Z,Ma H C,Zhang L.Cloverleaf interchange boundary extraction from airborne LiDAR data based on advanced neighborhood structure and contour analysis.Journal of Remote Sensing,2013,17(5):1146-1157.)
[4] Cheng G,Zhou P C,Yao X W,et al.Object detection in VHR optical remote sensing images via learning rotation-invariant HOG feature∥2016 4th International Workshop on Earth ObservationandRemoteSensingApplications(EORSA).Guangzhou,China:IEEE,2016:433-43.doi:10.1109/EORSA.2016.7552845.
[5] Cheng G,Han J W.A survey on object detection in optical remote sensing images.ISPRS Journal of Photogrammetry and Remote Sensing,2016,117:11-28.doi:10.1016/j.isprsjprs.2016.03.014.
[6] Liu W F,Zhang Z Q,Li S Y,et al.Road detectionbyusinga generalizedHough transform.Remote Sensing,2017,9(6):590.doi:10.3390/rs9060590.
[7] Bulley H N N,Bishop M P,Shroder J F,et al.Integration of classification tree analyses and spatial metrics to assess changes in supraglacial lakes in the Karakoram Himalaya.International Journal of Remote Sensing,2012,34(2):387-411.doi:10.1080/01431161.2012.705915.
[8] Holbling D,Füreder P,Antolini F,et al.A semi-automatedobject-basedapproachfor landslidedetectionvalidatedbypersistent scatterer interferometry measures and landslide inventories.Remote Sensing,2012,4(5):1310-1336.doi:10.3390/rs4051310.
[9] Ok A O.Automated detection of buildings from single VHR multispectral images using shadow information and graph cuts.ISPRS Journal of Photogrammetry and Remote Sensing,2013,86:21-40.doi:10.1016/j.isprsjprs.2013.09.004.
[10]Khadanga G,Jain K,Merugu S.Use of OBIA for extraction of cadastral parcels∥International Conference on Advances in Computing,Communications and Informatics(ICACCI).Jaipur,India:IEEE Press,2016:2226-2230.doi:10.1109/ICACCI.2016.7732382.
[11]Hosang J,Benenson R,Dollr P,et al.What makes for effective detection proposals?IEEE Transactions on Pattern Analysis and Machine Intelligence(TPAMI),2016,38(4):814-830.doi:10.1109/TPAMI.2015.2465908.
[12]Uijlings J R R,VanDeSande K E A,Gevers T,et al.Selective search for object recognition.International Journal of Computer Vision,2013,104(2):154-171.doi:10.1007/s11263-013-0620-5.
[13]Zhao C H,Liu W,Xu Y,et al.A spectral-spatial SVM-based multi-layer learning algorithm for hyperspectralimageclassification. Remote Sensing Letters,2018,9(3):218-227.doi:10.1080/2150704X.2017.1415475.
[14]Wu S L,Chen H D,Bai Y,et al.A remote sensing image classification method based on sparse representation. Multimedia Tools and Applications,2016,75(19):12137-12154.doi:10.1007/s11042-016-3320-7.
[15]Cheng G,Zhou P C,Han J W.Learning rotationinvariant convolutional neural networks for object detection in VHR optical remote sensing images.IEEE Transactions on Geoscience and Remote Sensing,2016,54(12):7405-7415.doi:10.1109/TGRS.2016.2601622.
[16]王小美,张红利.基于测地距离的KNN高光谱遥感图像分类.山西煤炭管理干部学院学报,2013,26(4):135-137.(Wang X M,Zhang H L.Hyperspectral remote sensing image classification using geodesic-based KNN.Journal of Shanxi Coalmining Administrators College,2013,26(4):135-137.)
[17]Wang J D,Zhang T,Song J K,et al.A survey on learning to hash.IEEE Transactions on Pattern Analysis and Machine Intelligence,2018,40(4):769-790.doi:10.1109/TPAMI.2017.2699960.
[18]Weiss Y,Torralba A,Fergus R.Spectral hashing∥International Conference on Neural InformationProcessingSystems(NIPS).Vancouver,Canada:Curran Associates Inc,2008,282(3):1753-1760.
[19]Jiang Q Y,Li W J.Scalable graph hashing with feature transformation∥Proceedings of the 24th International Conference on Artificial Intelligence(ICAI).Buenos Aires,Argentina:AAAI Press.2015:2248-2254.doi:10.1145/1814433.1814442.
[20]Shen F M,Shen C H,Shi Q F,et al.Inductive hashing on manifolds∥2013IEEE Conference on Computer Vision and Pattern Recognition.Portland,OR,USA:IEEE Press,2013:1562-1569.doi:10.1109/CVPR.2013.205.
[21]Liu W,Wang J,Ji R R,et al.Supervised hashing withkernels∥2012 IEEEConference onComputer Vision and Pattern Recognition.Providence,RI,USA:IEEE Press,2012:2074-2081.doi:10.1109/CVPR.2012.6247912.
[22]Norouzi M,Fleet D J.Minimal loss hashing for compact binary codes∥Proceedings of the 28th InternationalConferenceonInternational Conference on Machine Learning.Washington DC,USA:Omni Press,2011:353-360.
[23]Shen F M,Shen C H,Liu W,et al.Supervised discrete hashing∥2015 IEEE Conference on Computer VisionandPatternRecognition(CVPR).Boston,MA,USA:IEEE Press,2015:37-45.doi:10.1109/CVPR.2015.7298598.
[24]沈肖波.多视图嵌入学习方法及其应用研究.博士学位论文.南京:南京理工大学,2017.(Shen X B.Study of multi-view embedding learning techniques with applications.Ph.D.Dissertation.Nanjing:Nanjing University of Science and Technology,2017.)
[25]唐涛.图像仿射不变特征及其在遥感图像目标识别中的应用.硕士学位论文.长沙:国防科学技术大学,2006.(Tang T.Affine invariant feature and its application to target recognition in remote sensing images.Master Dissertation.Changsha:National University of Defense Technology,2006.)
[26]王凤国,冯象初,张小波.平稳小波变换在仿射不变性目标识别中的应用.计算机工程与应用,2007,43(21):239-241.(Wang F G,Feng X C,Zhang X B.Stationary wavelet transform for affineinvariantimageobjectrecognition.Computer Engineering and Applications,2007,43(21):239-241.)
[27]张洁玉,陈强,白小晶等.基于广义典型相关分析的仿射不变特征提取方法.电子与信息学报,2009,31(10):2465-2469.(Zhang J Y,Chen Q,Bai X J,et al.Affine invariant feature extraction algorithmbasedongeneralizedcanonical correlation analysis.Journal of Electronics&Information Technology,2009,31(10):2465-2469.)
[28]高峰,文项坚.利用仿射几何的仿射不变特征提取方法.中国图象图形学报,2011,16(3):389-397.(Gao F,Wen X J.A new method for affine invariants extraction based on affine geometry.Journal of Image and Graphics,2011,16(3):389-397.)
[29]Felzenszwalb P F,Huttenlocher D P.Efficient graph-based image segmentation.International Journal of Computer Vision,2004,59(2):167-181.doi:10.1023/b:visi.0000022288.19776.77.
[30]Long Y,Gong Y P,Xiao Z F,et al.Accurate object localization in remote sensing images based on convolutional neural networks.IEEE Transactions on Geoscience and Remote Sensing,2017,55(5):2486-2498.doi:10.1109/TGRS.2016.2645610.
[31]Felzenszwalb P F,Girshick R B,Mcallester D,et al.Object detection with discriminatively trained part-based models.IEEE Transactions on Pattern Analysis and Machine Intelligence,2010,32(9):1627-1645.doi:10.1109/TPAMI.2009.167.
[32]刘静,郭建,贺遵亮.基于Gist和PHOG特征的场景分类.计算机工程,2015,41(4):232-235,240.(Liu J,Guo J,He Z L.Scene classification based on gist and PHOG feature.Computer Engineering,2015,41(4):232-235,240.)