基于Hu不变矩特征的铁路轨道识别检测算法
|
摘要
针对当前铁路钢轨检测算法在识别中准确性和鲁棒性不高的问题,提出采用Hu不变矩特征实现轨道线搜索,并以B样条曲线为拟合模型的钢轨自动检测方法。算法根据视频帧中钢轨的边缘特征,通过改进的霍夫变换识别并确定图像空间的轨道线消隐边界,完成近远景区的标定。针对近景区直轨,通过直线模型拟合;在远景区,采用可漂移检测窗通过比对Hu不变矩来提取轨道特征点,以最小二乘法实现B样条曲线模型拟合。并制定模型更新和切换原则自动跟踪轨迹线。实验结果表明:轨道线平均跟踪时间为0. 081 s,可以提高钢轨检测识别的精确性和鲁棒性,能够更好地解决曲线轨道的模型拟合问题。
To address the problem of low accuracy and robustness in the existing railway track detection algorithm in recognition,an automatic rail inspection algorithm based on Hu invariant moments was proposed,which can be used to search the track line and which used B spline curve as the fitting model. Firstly,based on the edge feature of the rail in video frames,the algorithm was used to recognize the position of the track by the improved Hough transform,to determine the hidden boundary of the track line and to complete the calibration of the close and far range areas. For straight track in close range area,the linear model was used to calibrate the front straight rail. In the far range area,feature points were extracted by searching the Hu invariant moment feature based on the drift window search matching algorithm.Then,the least square method was used to fit B spline curve model. Finally,based on the model switching and the window searching strategy,the track line was tracked automatically. The experimental results show that the average time of tracking of the track line is 0. 081 s,and the algorithm can improve the accuracy and robustness of rail detection and better solve the problem of the model fitting of curved track.
To address the problem of low accuracy and robustness in the existing railway track detection algorithm in recognition,an automatic rail inspection algorithm based on Hu invariant moments was proposed,which can be used to search the track line and which used B spline curve as the fitting model. Firstly,based on the edge feature of the rail in video frames,the algorithm was used to recognize the position of the track by the improved Hough transform,to determine the hidden boundary of the track line and to complete the calibration of the close and far range areas. For straight track in close range area,the linear model was used to calibrate the front straight rail. In the far range area,feature points were extracted by searching the Hu invariant moment feature based on the drift window search matching algorithm.Then,the least square method was used to fit B spline curve model. Finally,based on the model switching and the window searching strategy,the track line was tracked automatically. The experimental results show that the average time of tracking of the track line is 0. 081 s,and the algorithm can improve the accuracy and robustness of rail detection and better solve the problem of the model fitting of curved track.
引文
[1]OH S,PARK S,JOUNG E. Approaches for Connection of Vision Based Monitoring System with Railway Signal System for Train Emergency Stop[C]//Proceedings of the IEEE International Conference on Information and Multimedia Technology.New York:IEEE,2009:59-63.
[2]PUNEKAR N S,RAUT A A. Improving Railway Safety with Obstacle Detection and Tracking System Using GPS-GSM Model[J]. International Journal of Science&Engineering Research,2013,4(8):288-292.
[3]肖衍,苏立勇.轨道交通全自动驾驶系统集成技术研究[J].中国铁路,2015,5(5):109-113.XIAO Yan,SU Liyong. Research on the Integrated Technology of the Automatic Driving System for Rail Transit[J]. China Railway,2015,5(5):109-113.
[4]刘大为,郭进,王小敏,等.中国铁路信号系统智能监测技术[J].西南交通大学学报,2014,49(9):904-912.LIU Dawei,GUO Jin,WANG Xiaomin,et al. Intelligent Monitoring Technologies for Railway Signaling Systems in China[J]. Journal of Southwest Jiaotong University,2014,49(9):904-912.
[5]KAZANSKIY N L,POPOV S B. Integrated Design Technology for Computer Vision Systems in Railway Transportation[J].Pattern Recognition and Image Analysis,2015,25(2):215-219.
[6]许华荣,王晓栋,方遒.基于B样条曲线模型的结构化道路检测算法[J].自动化学报,2011,37(3):270-275.XU Huarong,WANG Xiaodong,FANG Qiu. Structure Road Detection Algorithm Based on B-spline Curve Model[J]. Acta Automatica Sinica,2011,37(3):270-275.
[7]刘丽瑶,陈强,李艳娜,等.应用二阶Harris算子的高铁轨道近景影像特征点提取[J].测绘科学,2015,40(5):84-88.LIU Liyao,CHEN Qiang,LI Yanna,et al. Feature Points Extraction from High Speed Railway Track Images Based on Second Derivative Harris Operator[J]. Science of Surveying and Mapping,2015,40(5):84-88.
[8]李睿,武晓春.基于数字图像处理的直线型铁轨的自动识别[J].电视技术,2014,38(3):167-169.LI Rui,WU Xiaochun. Automatic Identify of Linear Tracks Based on Digital Image Processing[J]. Video Engineering,2014,38(3):167-169.
[9]居耀勇,陈黎.基于主方向迭代校正的铁轨检测算法[J].武汉科技大学学报,2012,35(6):473-476.JU Yaoyong,CHEN Li. An Iterative Principal Orientation Adjustment-based Algorithm for Railway Track Detection[J].Journal of Wuhan University of Science and Technology,2012,35(6):473-476.
[10]超木日力格.机车司机视野扩展系统及路轨障碍物检测的研究[D].北京:北京交通大学,2012:23-42.
[11]王前选,梁习锋,刘应龙,等.铁路钢轨视觉识别检测方法[J]:中南大学学报(自然科学版),2014,45(7):2496-2502.WANG Qianxuan,LIANG Xifeng,LIU Yinglong,et al. Railway Rail Identification Detection Method Using Machine Vision[J]. Journal of Central South University(Science and Technology),2014,45(7):2496-2502.
[12]CHEN L J,WANG K Q,XIE Y H,et al. The Segmentation of Timber Defects Based on Color and the Mathematical Morphology[J]. Optik,2014,125(3):965-967.
[13]易思蓉,聂良涛,秦方方.基于动力学分析的高速铁路最小曲线半径研究[J].西南交通大学学报,2013,48(1):16-20.YI Sirong,NIE Liangtao,QIN Fangfang. Research on the Minimum Curve Radius of High-speed Railway Based on Dynamic Analysis[J]. Journal of Southwest Jiaotong University,2013,48(1):16-20.
[14]CHEN Q,WANG H. A Real-time Lane Detection Algorithm Based on a Hyperbola-pair Model[C]//IEEE Intelligent Vehicles Symposium 2006. New York:IEEE,2006:510-515.
[15]LIM K H,SENG K P,ANG L M. River Flow Lane Detection and Kalman Filtering-based B-spline Lane Tracking[J]. International Journal of Vehicular Technology,2012,2012(9):1-10.
[16]鲁曼,蔡自兴,李仪.道路区域分割的车道线检测方法[J].智能系统学报,2010,5(6):505-509.LU Man,CAI Zixing,LI Yi. A Lane Detection Method Based on Road Segmentation[J]. CAAI Transactions on Intelligent Systems,2010,5(6):505-509.
[17]GONZALEZ R C,WOODS R E,EDDINS S L. Digital Image Processing Using MATLAB[M].北京:电子工业出版社,2012:296-302.
[18]史红梅,柴华,王尧,等.基于目标识别与跟踪的嵌入式铁路异物侵限检测算法研究[J].铁道学报,2015,37(7):58-65.SHI Hongmei,CHAI Hua,WANG Yao,et al. Study on Railway Embedded Detection Algorithm for Railway Intrusion Based on Object Recognition and Tracking[J]. Journal of the China Railway Society,2015,37(7):58-65.
[2]PUNEKAR N S,RAUT A A. Improving Railway Safety with Obstacle Detection and Tracking System Using GPS-GSM Model[J]. International Journal of Science&Engineering Research,2013,4(8):288-292.
[3]肖衍,苏立勇.轨道交通全自动驾驶系统集成技术研究[J].中国铁路,2015,5(5):109-113.XIAO Yan,SU Liyong. Research on the Integrated Technology of the Automatic Driving System for Rail Transit[J]. China Railway,2015,5(5):109-113.
[4]刘大为,郭进,王小敏,等.中国铁路信号系统智能监测技术[J].西南交通大学学报,2014,49(9):904-912.LIU Dawei,GUO Jin,WANG Xiaomin,et al. Intelligent Monitoring Technologies for Railway Signaling Systems in China[J]. Journal of Southwest Jiaotong University,2014,49(9):904-912.
[5]KAZANSKIY N L,POPOV S B. Integrated Design Technology for Computer Vision Systems in Railway Transportation[J].Pattern Recognition and Image Analysis,2015,25(2):215-219.
[6]许华荣,王晓栋,方遒.基于B样条曲线模型的结构化道路检测算法[J].自动化学报,2011,37(3):270-275.XU Huarong,WANG Xiaodong,FANG Qiu. Structure Road Detection Algorithm Based on B-spline Curve Model[J]. Acta Automatica Sinica,2011,37(3):270-275.
[7]刘丽瑶,陈强,李艳娜,等.应用二阶Harris算子的高铁轨道近景影像特征点提取[J].测绘科学,2015,40(5):84-88.LIU Liyao,CHEN Qiang,LI Yanna,et al. Feature Points Extraction from High Speed Railway Track Images Based on Second Derivative Harris Operator[J]. Science of Surveying and Mapping,2015,40(5):84-88.
[8]李睿,武晓春.基于数字图像处理的直线型铁轨的自动识别[J].电视技术,2014,38(3):167-169.LI Rui,WU Xiaochun. Automatic Identify of Linear Tracks Based on Digital Image Processing[J]. Video Engineering,2014,38(3):167-169.
[9]居耀勇,陈黎.基于主方向迭代校正的铁轨检测算法[J].武汉科技大学学报,2012,35(6):473-476.JU Yaoyong,CHEN Li. An Iterative Principal Orientation Adjustment-based Algorithm for Railway Track Detection[J].Journal of Wuhan University of Science and Technology,2012,35(6):473-476.
[10]超木日力格.机车司机视野扩展系统及路轨障碍物检测的研究[D].北京:北京交通大学,2012:23-42.
[11]王前选,梁习锋,刘应龙,等.铁路钢轨视觉识别检测方法[J]:中南大学学报(自然科学版),2014,45(7):2496-2502.WANG Qianxuan,LIANG Xifeng,LIU Yinglong,et al. Railway Rail Identification Detection Method Using Machine Vision[J]. Journal of Central South University(Science and Technology),2014,45(7):2496-2502.
[12]CHEN L J,WANG K Q,XIE Y H,et al. The Segmentation of Timber Defects Based on Color and the Mathematical Morphology[J]. Optik,2014,125(3):965-967.
[13]易思蓉,聂良涛,秦方方.基于动力学分析的高速铁路最小曲线半径研究[J].西南交通大学学报,2013,48(1):16-20.YI Sirong,NIE Liangtao,QIN Fangfang. Research on the Minimum Curve Radius of High-speed Railway Based on Dynamic Analysis[J]. Journal of Southwest Jiaotong University,2013,48(1):16-20.
[14]CHEN Q,WANG H. A Real-time Lane Detection Algorithm Based on a Hyperbola-pair Model[C]//IEEE Intelligent Vehicles Symposium 2006. New York:IEEE,2006:510-515.
[15]LIM K H,SENG K P,ANG L M. River Flow Lane Detection and Kalman Filtering-based B-spline Lane Tracking[J]. International Journal of Vehicular Technology,2012,2012(9):1-10.
[16]鲁曼,蔡自兴,李仪.道路区域分割的车道线检测方法[J].智能系统学报,2010,5(6):505-509.LU Man,CAI Zixing,LI Yi. A Lane Detection Method Based on Road Segmentation[J]. CAAI Transactions on Intelligent Systems,2010,5(6):505-509.
[17]GONZALEZ R C,WOODS R E,EDDINS S L. Digital Image Processing Using MATLAB[M].北京:电子工业出版社,2012:296-302.
[18]史红梅,柴华,王尧,等.基于目标识别与跟踪的嵌入式铁路异物侵限检测算法研究[J].铁道学报,2015,37(7):58-65.SHI Hongmei,CHAI Hua,WANG Yao,et al. Study on Railway Embedded Detection Algorithm for Railway Intrusion Based on Object Recognition and Tracking[J]. Journal of the China Railway Society,2015,37(7):58-65.