基于随机投影深度函数的停车场车辆提取方法
|
摘要
为精确提取露天停车场内颜色混杂的车辆,提出一种基于随机投影深度函数的车辆提取方法.随机投影深度函数可有效区分RGB彩色空间中数据集的数据中心与离群值,充分利用各车辆颜色特征的复杂性及其与停车场背景颜色特征的差异性,凸显具有离群值颜色特征的车辆.首先,利用随机投影深度函数对彩色遥感影像中各像素颜色特征进行排序得到深度场影像;然后,对深度场影像做形态学闭运算并选取合适的随机投影深度值作为阈值,二值化闭运算后的深度场影像,实现车辆初始提取;最后,结合决策树分析与形态学运算实现车辆精确提取.实验结果表明,随机投影深度函数可有效处理彩色遥感影像中各种颜色车辆所表现的"同物异谱"现象,在深度场影像中凸显不同颜色的车辆,有效提高车辆提取效率;辅助以简单的后处理可实现遥感影像中不同场景停车场车辆提取.
An algorithm of extraction of vehicles based on random projection depth function is proposed for accurately extracting vehicles with different colors in outdoor parking lots. The random projection depth function can effectively distinguish the center and outlier of the data set in RGB color space, and in this way, the vehicles whose color characteristics act as outlier are highlighted. First, the random projection depth function is used to sort the color characteristic of each pixel to obtain random projection depth value,f orming the depth field image; Then,morphological closed operation is carried out for the depth field image, and an appropriate random projection depth value is selected as the threshold to binarize the image;Finally, the vehicles are accurately extracted from the parking lot by decision tree algorithm and morphological operations.The experimental results show that the random projection depth function can effectively deal with the " same body with different spectrum" phenomenon of various color vehicles in remote sensing images. The vehicles of different colors are highlighted in the depth field image,which can effectively improve the efficiency of vehicle extraction, and extraction of vehicles from parking lot of remote sensing images can be realized accurately by combining random projection depth function and simple post-processing.
An algorithm of extraction of vehicles based on random projection depth function is proposed for accurately extracting vehicles with different colors in outdoor parking lots. The random projection depth function can effectively distinguish the center and outlier of the data set in RGB color space, and in this way, the vehicles whose color characteristics act as outlier are highlighted. First, the random projection depth function is used to sort the color characteristic of each pixel to obtain random projection depth value,f orming the depth field image; Then,morphological closed operation is carried out for the depth field image, and an appropriate random projection depth value is selected as the threshold to binarize the image;Finally, the vehicles are accurately extracted from the parking lot by decision tree algorithm and morphological operations.The experimental results show that the random projection depth function can effectively deal with the " same body with different spectrum" phenomenon of various color vehicles in remote sensing images. The vehicles of different colors are highlighted in the depth field image,which can effectively improve the efficiency of vehicle extraction, and extraction of vehicles from parking lot of remote sensing images can be realized accurately by combining random projection depth function and simple post-processing.
引文
[1] Jens Leitloff,Dominik Rosenbaum,Franz Kurz,OliverM eynberg Peter Reinartz. An operational system for esti-mating road traffic information from aerial images[J]. Re-mote Sensing,2014,6(11):11315-11341.
[2]Aleena Ajay,V Sowmya,K P Soman. Vehicle detection inaerial imagery using eigen features[A]. Proceedings ofIEEE International Conference on Communication and Sig-nal Processing[C]. Trabzon,Turkey:IEEE,2017. 6-8.
[3] Lihong Wan,Laiwen Zheng,Hong Huo,Tao Fang. Affineinvariant description and large-margin dimensionality re-duction for target detection in optical remote sensing ima-ges[J]. IEEE Geoscience and Remote Sensing Letters,2017,14(7):1116-1120.
[4] Yong Tang,Congzhe Zhang,Renshu Gu,Peng Li,BinYang. Vehicle detection and recognition for intelligent traf-fic surveillance system[J]. M ultimedia tools and applica-tions,2017,76(4):5817-5832.
[5] Minkyu Cheon,Wonju Lee,Changyong Yoon,MignonPark. Vision-based vehicle detection system w ith consider-ation of the detecting location[J]. IEEE Transactions onIntelligent Transportation Systems,2012,3(13):1243-1252.
[6]Zehang Sun,George Bebis,Ronald Miller. On-road vehicledetection using evolutionary Gabor filter optimization[J].IEEE Transactions on Intelligent Transportation Systems,2005,6(2):125-137.
[7]Sheng Wang. Vehicle detection on aerial images by extrac-ting corner features for rotational invariant shape matching[A]. Proceedings of IEEE International Conference onComputer and Information Technology[C]. Pafos,Cyprus:IEEE,2011. 171-175.
[8] Binfeng Lin,Yiming Chan,Lichen Fu,Peiyung Hsiao. In-tegrating appearance and edge features for sedan vehicledetection in the blind-spot area[J]. IEEE Transactions onIntelligent Transportation Systems,2012,13(2):737-747.
[9] Dong Guo,Ming Xie. Color modeling by spherical influ-ence field in sensing driving environment[A]. Proceedingsof the IEEE Intelligent Vehicles Symposium[C]. Dear-born,USA:IEEE,2000. 249-254.
[10]Juan Rojas,Dan Crisman. Vehicle detection in colour ima-ges[A]. Proceedings of IEEE Conference IntelligentTransportation System[C]. Boston,USA:IEEE,1997.403-408.
[11] Yijun Zuo,Robert Serfling. General notions of statisticaldepth function[J]. The Annals of Statistics,2000,28(2):461-482.
[12]Yijun Zuo. Projection-based depth function and associatedmedians[J]. The Annals of Statistics,2003,31(5):1460-1490.
[13]Velasco Santiago,Angulo Jesus. Random projection depthfor multivariate mathematical morphology[J]. IEEE Jour-nal of Selected Topics in Signal Processing,2012,6(7):753-763.
[14]Peter Rousseeuw,Ida Ruts,John Tukey. The bagplot:a bi-variate boxplot[J]. The American Statistician,1999,53(2):382-387.
[15]Xiaohui Liu,Yijun Zuo. Computing projection depth andits associated estimators[J]. Statistics&Computing,2014,24(1):51-63.
[16]许新征,丁世飞,史忠植,贾伟宽.图像分割的新理论和新方法[J].电子学报,2010,38(2A):76-82.Xu Xinzheng,Ding Shifei,Shi Zhongzhi,Jia Weikuan.New theories and methods of image segmentation[J]. Ac-ta Electronica Sinica,2010,38(2A):76-82.(in Chi-nese)
[17]孔斌.快速连通域分析算法及实现[J].模式识别与人工智能,2003,16(1):110-115.Kong Bin. A fast connected component algorithm and itsimplementation[J]. Pattern Recognition and Artificial In-telligence,16(1):110-115.(in Chinese)
[18]魏长安,姜守达.基于形态重构与跟踪的红外小目标检测算法[J].电子学报,2009,37(4):850-853.Wei Changan,Jiang Shouda. Infrared smalltarget detectionalgorithm based on morphological reconstruction operatorand tracking[J]. Acta Electronica Sinica,2009,37(4):850-853.(in Chinese)
[19]Zezhong Zheng,Guoqing Zhou,Yong Wang. A novel ve-hicle detection method w ith high resoluti on highw ay aeri-al image[J]. IEEE Journal of Selected Topics in AppliedEarth Observations and Remote Sensing,2013,6(6):2338-2343.
[20] Hakki Karaimer,Yalin Bactanlar. Car detection with om-nidirectional cameras using Haar-like features and casca-ded boosting[A]. Proceedings of the Conference on Sig-nal Processing and Communications Applications[C].Hangzhou,China:IEEE,2017. 11-13.
[21] Jaeyoung Choi,Youngkyu Yang. Vehicle detection fromaerial images using local shape information[A]. Ad-vances in Image and Video Technology[C]. Berlin,Hei-delberg:Springer,2009. 227-236.
[23]Xutong Niu. A semi-automatic framework for highway ex-traction and vehicle detection based on algometric deform-able model[J]. ISPRS Journal of Photogrammetry&Re-mote Sensing,2006,61(3-4):170-186.
[2]Aleena Ajay,V Sowmya,K P Soman. Vehicle detection inaerial imagery using eigen features[A]. Proceedings ofIEEE International Conference on Communication and Sig-nal Processing[C]. Trabzon,Turkey:IEEE,2017. 6-8.
[3] Lihong Wan,Laiwen Zheng,Hong Huo,Tao Fang. Affineinvariant description and large-margin dimensionality re-duction for target detection in optical remote sensing ima-ges[J]. IEEE Geoscience and Remote Sensing Letters,2017,14(7):1116-1120.
[4] Yong Tang,Congzhe Zhang,Renshu Gu,Peng Li,BinYang. Vehicle detection and recognition for intelligent traf-fic surveillance system[J]. M ultimedia tools and applica-tions,2017,76(4):5817-5832.
[5] Minkyu Cheon,Wonju Lee,Changyong Yoon,MignonPark. Vision-based vehicle detection system w ith consider-ation of the detecting location[J]. IEEE Transactions onIntelligent Transportation Systems,2012,3(13):1243-1252.
[6]Zehang Sun,George Bebis,Ronald Miller. On-road vehicledetection using evolutionary Gabor filter optimization[J].IEEE Transactions on Intelligent Transportation Systems,2005,6(2):125-137.
[7]Sheng Wang. Vehicle detection on aerial images by extrac-ting corner features for rotational invariant shape matching[A]. Proceedings of IEEE International Conference onComputer and Information Technology[C]. Pafos,Cyprus:IEEE,2011. 171-175.
[8] Binfeng Lin,Yiming Chan,Lichen Fu,Peiyung Hsiao. In-tegrating appearance and edge features for sedan vehicledetection in the blind-spot area[J]. IEEE Transactions onIntelligent Transportation Systems,2012,13(2):737-747.
[9] Dong Guo,Ming Xie. Color modeling by spherical influ-ence field in sensing driving environment[A]. Proceedingsof the IEEE Intelligent Vehicles Symposium[C]. Dear-born,USA:IEEE,2000. 249-254.
[10]Juan Rojas,Dan Crisman. Vehicle detection in colour ima-ges[A]. Proceedings of IEEE Conference IntelligentTransportation System[C]. Boston,USA:IEEE,1997.403-408.
[11] Yijun Zuo,Robert Serfling. General notions of statisticaldepth function[J]. The Annals of Statistics,2000,28(2):461-482.
[12]Yijun Zuo. Projection-based depth function and associatedmedians[J]. The Annals of Statistics,2003,31(5):1460-1490.
[13]Velasco Santiago,Angulo Jesus. Random projection depthfor multivariate mathematical morphology[J]. IEEE Jour-nal of Selected Topics in Signal Processing,2012,6(7):753-763.
[14]Peter Rousseeuw,Ida Ruts,John Tukey. The bagplot:a bi-variate boxplot[J]. The American Statistician,1999,53(2):382-387.
[15]Xiaohui Liu,Yijun Zuo. Computing projection depth andits associated estimators[J]. Statistics&Computing,2014,24(1):51-63.
[16]许新征,丁世飞,史忠植,贾伟宽.图像分割的新理论和新方法[J].电子学报,2010,38(2A):76-82.Xu Xinzheng,Ding Shifei,Shi Zhongzhi,Jia Weikuan.New theories and methods of image segmentation[J]. Ac-ta Electronica Sinica,2010,38(2A):76-82.(in Chi-nese)
[17]孔斌.快速连通域分析算法及实现[J].模式识别与人工智能,2003,16(1):110-115.Kong Bin. A fast connected component algorithm and itsimplementation[J]. Pattern Recognition and Artificial In-telligence,16(1):110-115.(in Chinese)
[18]魏长安,姜守达.基于形态重构与跟踪的红外小目标检测算法[J].电子学报,2009,37(4):850-853.Wei Changan,Jiang Shouda. Infrared smalltarget detectionalgorithm based on morphological reconstruction operatorand tracking[J]. Acta Electronica Sinica,2009,37(4):850-853.(in Chinese)
[19]Zezhong Zheng,Guoqing Zhou,Yong Wang. A novel ve-hicle detection method w ith high resoluti on highw ay aeri-al image[J]. IEEE Journal of Selected Topics in AppliedEarth Observations and Remote Sensing,2013,6(6):2338-2343.
[20] Hakki Karaimer,Yalin Bactanlar. Car detection with om-nidirectional cameras using Haar-like features and casca-ded boosting[A]. Proceedings of the Conference on Sig-nal Processing and Communications Applications[C].Hangzhou,China:IEEE,2017. 11-13.
[21] Jaeyoung Choi,Youngkyu Yang. Vehicle detection fromaerial images using local shape information[A]. Ad-vances in Image and Video Technology[C]. Berlin,Hei-delberg:Springer,2009. 227-236.
[23]Xutong Niu. A semi-automatic framework for highway ex-traction and vehicle detection based on algometric deform-able model[J]. ISPRS Journal of Photogrammetry&Re-mote Sensing,2006,61(3-4):170-186.