一种近景大斜角的限界检测相机标定方法
|
摘要
研究一种基于单目视觉的非接触式站台限界检测系统,并针对近景大范围下相机标定的问题,提出一种基于大型点阵标定板的新的综合标定方法。提取有效标定区域进行预处理,并提出一种局部阈值迭代分割的方法分割出基准圆,然后提取圆心坐标并编码,最后利用透视变换求出最优变换矩阵进行图像矫正还原,像素标定误差可达到2pixel以内。经在上海地铁的现场测试验证,本文方法标定后的检测系统可准确获取站台断面轮廓几何尺寸,多次反复测量后轨距精度在±1 mm以内,站台高度和站台距离精度可达到±5 mm以内。
In this paper, a non-contact platform limit detection system based on monocular vision was studied, and a new comprehensive calibration method based on large-scale lattice calibration plate was proposed for the problem of camera calibration under a wide range of close-range vision. The effective calibration region was extracted for preprocessing. And this paper propose a local threshold iterative segmentation method to segment the reference circle, then extracts the center coordinates and encode, and finally uses the perspective transformation to find the optimal transformation matrix for image correction and reduction, pixel calibration error can be Within 2 pixel. After field tests in the Shanghai subway, the detection system after calibration of this method can accurately obtain the geometrical dimensions of the platform section profile. After repeated measurements, the gauge accuracy is within ± 1 mm, and the platform height and platform distance accuracy can be within ±5 mm.
In this paper, a non-contact platform limit detection system based on monocular vision was studied, and a new comprehensive calibration method based on large-scale lattice calibration plate was proposed for the problem of camera calibration under a wide range of close-range vision. The effective calibration region was extracted for preprocessing. And this paper propose a local threshold iterative segmentation method to segment the reference circle, then extracts the center coordinates and encode, and finally uses the perspective transformation to find the optimal transformation matrix for image correction and reduction, pixel calibration error can be Within 2 pixel. After field tests in the Shanghai subway, the detection system after calibration of this method can accurately obtain the geometrical dimensions of the platform section profile. After repeated measurements, the gauge accuracy is within ± 1 mm, and the platform height and platform distance accuracy can be within ±5 mm.
引文
[1]李甍,于龙,张冬凯.地铁设备限界随机因素分析[J].铁道科学与工程学报, 2015, 12(3):657-662.LI Wei, YU Long, ZHANG Dongkai. Analysis on random factors of subway equipment boundary[J]. Journal of Railway Science and Engineering, 2015, 12(3):657-662.
[2]史红梅,吴攀,余祖俊,等.基于点集配准的非接触式地铁限界检测系统标定方法研究[J].铁道学报, 2014,36(11):16-22.SHI Hongmei, WU Pan, YU Zujun, et al. Study on calibration method of contactless metro boundary detection system based on point set registration[J].Journal of the China Railway Society, 2014, 36(11):16-22.
[3]钱丽丹,杨军伟.基于精密运动的激光平面标定方法[J].计算机工程, 2013, 39(4):318-321.QIAN Lidan, YANG Junwei. Laser plane calibration method based on precise motion[J]. Computer Engineering, 2013, 39(4):318-321.
[4]吴凡路,刘建军,任鑫,等.基于圆形标志点的深空探测全景相机标定方法[J].光学学报, 2013, 33(11):147-153.WU Fanlu, LIU Jianjun, REN Xin, et al. Calibration method of deep space exploration panoramic camera based on circular landmarks[J]. Journal of Optics, 2013,33(11):147-153.
[5] ZHANG Z Y. A flexible new technique for camera calibration[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(11):1330-1334.
[6]卜鹏辉,赵宏,谷飞飞,等.一种基于平面靶标的全自动标定方法[J].光学学报, 2014, 34(10):200-204.BU Penghui, ZHAO Hong, GU Feifei, et al. A fully automatic calibration method based on planar targets[J].Acta Optica Sinica, 2014, 34(10):200-204.
[7]刘震,尚砚娜.立体视觉传感器的一种灵活靶标新标定方法[J].光学学报, 2013, 33(3):205-211.LIU Zhen, SHANG Yanna. A new calibration method for flexible targets of stereo vision sensors[J]. Acta Optica Sinica, 2013, 33(3):205-211.
[8]梁力,尹东斐,王川.高精度摄像机标定模板的设计及识别算法[J].西安交通大学学报, 2011, 45(4):82-85.LIANG Li, YIN Dongfei, WANG Chuan. Design and recognition algorithm of high-precision camera calibration template[J]. Journal of Xi’an Jiaotong University, 2011, 45(4):82-85.
[9]夏仁波,刘伟军,赵吉宾,等.基于圆形标志点的全自动相机标定方法[J].仪器仪表学报, 2009, 30(2):368-373.XIA Renbo, LIU Weijun, ZHAO Jibin, et al. Automatic camera calibration method based on circular marker[J].Chinese Journal of Scientific Instrument, 2009, 30(2):368-373.
[10]胡浩,梁晋,唐正宗,等.大视场多像机视频测量系统的全局标定[J].光学精密工程, 2012, 20(2):369-378.HU Hao, LIANG Jin, TANG Zhengzong, et al. Global calibration of video camera measurement system for large field of view[J]. Optics and Precision Engineering, 2012,20(2):369-378.
[11]余佳磊,何越磊,龚佩毅,等.基于区域生长重心的大型多阵列圆标定板基准点的提取技术[J].计算机测量与控制, 2017, 25(11):146-149.YU Jialei, HE Yuelei, GONG Peiyi, et al. Extraction techniques of large-scale multi-array fixed plate reference points based on regional growing gravity[J]. Computer Measurement&Control, 2017, 25(11):146-149.
[12] Flusser J, Farokhi S, Hoschl C, et al. Recognition of images degraded by Gaussian blur[J]. IEEE transactions on image processing:a publication of the IEEE Signal Processing Society, 2016, 25(2):88-99.
[13] Sezgin M, Sankur B. Survey over image thresholding techniques and quantitative performance evaluation[J].Journal of Electronic Imaging, 2004, 13(1):146-168.
[14]闫建平,梁强,李尊芝,等.连通域标识法在FMI图像溶洞信息定量拾取中的应用[J].地球物理学报, 2016,59(12):4759-4770.YAN Jianping, LIANG Qiang, LI Zunzhi, et al.Application of connected domain identification method in quantitative pickup of karst information in fmi images[J].Chinese Journal of Geophysics, 2016, 59(12):4759-4770.
[15]蔡国榕,李绍滋,吴云东,等.一种透视不变的图像匹配算法[J].自动化学报, 2013, 39(7):1053-1061.CAI Guorong, LI Shaozi, WU Yundong, et al. A perspective-invariant image matching algorithm[J]. Acta Automatica Sinica, 2013, 39(7):1053-1061.
[2]史红梅,吴攀,余祖俊,等.基于点集配准的非接触式地铁限界检测系统标定方法研究[J].铁道学报, 2014,36(11):16-22.SHI Hongmei, WU Pan, YU Zujun, et al. Study on calibration method of contactless metro boundary detection system based on point set registration[J].Journal of the China Railway Society, 2014, 36(11):16-22.
[3]钱丽丹,杨军伟.基于精密运动的激光平面标定方法[J].计算机工程, 2013, 39(4):318-321.QIAN Lidan, YANG Junwei. Laser plane calibration method based on precise motion[J]. Computer Engineering, 2013, 39(4):318-321.
[4]吴凡路,刘建军,任鑫,等.基于圆形标志点的深空探测全景相机标定方法[J].光学学报, 2013, 33(11):147-153.WU Fanlu, LIU Jianjun, REN Xin, et al. Calibration method of deep space exploration panoramic camera based on circular landmarks[J]. Journal of Optics, 2013,33(11):147-153.
[5] ZHANG Z Y. A flexible new technique for camera calibration[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(11):1330-1334.
[6]卜鹏辉,赵宏,谷飞飞,等.一种基于平面靶标的全自动标定方法[J].光学学报, 2014, 34(10):200-204.BU Penghui, ZHAO Hong, GU Feifei, et al. A fully automatic calibration method based on planar targets[J].Acta Optica Sinica, 2014, 34(10):200-204.
[7]刘震,尚砚娜.立体视觉传感器的一种灵活靶标新标定方法[J].光学学报, 2013, 33(3):205-211.LIU Zhen, SHANG Yanna. A new calibration method for flexible targets of stereo vision sensors[J]. Acta Optica Sinica, 2013, 33(3):205-211.
[8]梁力,尹东斐,王川.高精度摄像机标定模板的设计及识别算法[J].西安交通大学学报, 2011, 45(4):82-85.LIANG Li, YIN Dongfei, WANG Chuan. Design and recognition algorithm of high-precision camera calibration template[J]. Journal of Xi’an Jiaotong University, 2011, 45(4):82-85.
[9]夏仁波,刘伟军,赵吉宾,等.基于圆形标志点的全自动相机标定方法[J].仪器仪表学报, 2009, 30(2):368-373.XIA Renbo, LIU Weijun, ZHAO Jibin, et al. Automatic camera calibration method based on circular marker[J].Chinese Journal of Scientific Instrument, 2009, 30(2):368-373.
[10]胡浩,梁晋,唐正宗,等.大视场多像机视频测量系统的全局标定[J].光学精密工程, 2012, 20(2):369-378.HU Hao, LIANG Jin, TANG Zhengzong, et al. Global calibration of video camera measurement system for large field of view[J]. Optics and Precision Engineering, 2012,20(2):369-378.
[11]余佳磊,何越磊,龚佩毅,等.基于区域生长重心的大型多阵列圆标定板基准点的提取技术[J].计算机测量与控制, 2017, 25(11):146-149.YU Jialei, HE Yuelei, GONG Peiyi, et al. Extraction techniques of large-scale multi-array fixed plate reference points based on regional growing gravity[J]. Computer Measurement&Control, 2017, 25(11):146-149.
[12] Flusser J, Farokhi S, Hoschl C, et al. Recognition of images degraded by Gaussian blur[J]. IEEE transactions on image processing:a publication of the IEEE Signal Processing Society, 2016, 25(2):88-99.
[13] Sezgin M, Sankur B. Survey over image thresholding techniques and quantitative performance evaluation[J].Journal of Electronic Imaging, 2004, 13(1):146-168.
[14]闫建平,梁强,李尊芝,等.连通域标识法在FMI图像溶洞信息定量拾取中的应用[J].地球物理学报, 2016,59(12):4759-4770.YAN Jianping, LIANG Qiang, LI Zunzhi, et al.Application of connected domain identification method in quantitative pickup of karst information in fmi images[J].Chinese Journal of Geophysics, 2016, 59(12):4759-4770.
[15]蔡国榕,李绍滋,吴云东,等.一种透视不变的图像匹配算法[J].自动化学报, 2013, 39(7):1053-1061.CAI Guorong, LI Shaozi, WU Yundong, et al. A perspective-invariant image matching algorithm[J]. Acta Automatica Sinica, 2013, 39(7):1053-1061.