基于特征点提取技术的车辆行车视距检测算法研究
|
摘要
为了提高弯道行车视距检测的有效性,采用中值滤波及Sobel算子对道路图像进行车道线边缘提取,接着对图像进行二值化处理,基于车道标识线固有特征对车道标识线进行筛选并提取计算行车视距所需要特征点的坐标值,然后利用小孔成像原理建立驾驶员行车视距测距模型。为验证本研究所提出算法的正确性及有效性,设计并实施了道路图像行车视距离线测算实验,通过对比实验数据可以得出:所提出的行车视距检测模型绝对误差平均值为1. 3 m,相对误差平均值为5. 3%,能够较好地为车辆辅助驾驶系统的行车视距功能开发提供算法支撑,提高车辆弯道行车的安全性。
In order to improve the effectiveness of the driving sight distance detection of curved roads,the median filtering algorithm and Sobel operator was used to extract the lane line edge of the road image in this paper,and then the image was binarized. Based on the inherent characteristics of the lane marking line,the line was screened and the coordinate values of the feature points for calculating the driving sight distance were extracted,then the driver's driving sight distance measurement model was established by using the small hole imaging principle. In order to verify the correctness and effectiveness of the proposed algorithm,the driving sight distance measurement experiment through road image is designed and implemented. The experimental result shows that the absolute error of the driving sight distance detection model proposed in this paper is 1. 3 m. The average relative error is 5.3%,which can provide algorithm support for the development of the driving sight distance function of the vehicle assisted driving system,and improve the driving safety on curves.
In order to improve the effectiveness of the driving sight distance detection of curved roads,the median filtering algorithm and Sobel operator was used to extract the lane line edge of the road image in this paper,and then the image was binarized. Based on the inherent characteristics of the lane marking line,the line was screened and the coordinate values of the feature points for calculating the driving sight distance were extracted,then the driver's driving sight distance measurement model was established by using the small hole imaging principle. In order to verify the correctness and effectiveness of the proposed algorithm,the driving sight distance measurement experiment through road image is designed and implemented. The experimental result shows that the absolute error of the driving sight distance detection model proposed in this paper is 1. 3 m. The average relative error is 5.3%,which can provide algorithm support for the development of the driving sight distance function of the vehicle assisted driving system,and improve the driving safety on curves.
引文
[1]王磊,吕璞,林永杰.高速公路交通事故影响因素分析及伤害估计[J].中国安全科学学报,2016,26(3):86-90.
[2] ISMAIL K,SAYED T. New algorithm for calculating 3D available sight distance[J]. Journal Transportation Engineering,2007,133(10):572-581.
[3] YASSER H. Effect of driver and road characteristics on required preview sight distance[J]. Canadian Journal of Civil Engineering,2002,29(2):276-288.
[4]史颖,邓建华,骆豪.基于视频检测的车辆变道轨迹识别方法研究[J].交通科技与经济,2017,19(3):48-53.
[5]曹?,李进,张智欣.机器视觉辅助驾驶系统的车辆检测和测距[J].电子技术与软件工程,2018(17):238.
[6]杨帆,张皓,马新文,等.基于FPGA的图像处理系统[J].华中科技大学学报(自然科学版),2015,43(2):119-123.
[7]王宝锋,齐志权,马国成,等.基于动态区域规划的双模型车道线识别方法[J].北京理工大学学报,2014,34(5):485-489.
[8]梁朱冬,陈洪洋.基于改进最小二乘法车道线模型的车道标识线检测识别[J].汽车实用技术,2017(15):98-100.
[9]刘向阳,周淼,郭腾峰.公路行车视距分析与验证[J].中国公路学报,2010,23(S1):36-41.
[10]刘永涛.基于环境感知技术的客运车辆危险行驶状态辨识技术研究[D].西安:长安大学,2015.
[11]傅思勇,吴禄慎,陈华伟,等.综合多畸变因素的摄像机标定[J].仪器仪表学报,2018,39(2):248-256.
[12]陈雨人,付云天,汪凡.基于支持向量回归的视距计算模型建立和应用[J].中国公路学报,2018,31(4):105-113.
[13]李璐璐,赵文川,伍凡,等.摄像机标定中的特征点提取算法研究与改进[J].光学学报,2014,34(5):179-186.
[14]董斌,刘谦.基于三维模型的高速公路中央分隔带外侧车道视距计算方法研究[J].交通科技,2017(1):148-150.
[15]文涛,左东广,李忠科,等.一种鲁棒的高精度摄像机标定方法[J].计算机应用研究,2015,32(11):3489-3491,3495.
[16] ZHANG Z. A flexible new technique for camera calibration[J]. IEEE Transactions on pattern analysis and machine intelligence,2000,22:110-119.
[2] ISMAIL K,SAYED T. New algorithm for calculating 3D available sight distance[J]. Journal Transportation Engineering,2007,133(10):572-581.
[3] YASSER H. Effect of driver and road characteristics on required preview sight distance[J]. Canadian Journal of Civil Engineering,2002,29(2):276-288.
[4]史颖,邓建华,骆豪.基于视频检测的车辆变道轨迹识别方法研究[J].交通科技与经济,2017,19(3):48-53.
[5]曹?,李进,张智欣.机器视觉辅助驾驶系统的车辆检测和测距[J].电子技术与软件工程,2018(17):238.
[6]杨帆,张皓,马新文,等.基于FPGA的图像处理系统[J].华中科技大学学报(自然科学版),2015,43(2):119-123.
[7]王宝锋,齐志权,马国成,等.基于动态区域规划的双模型车道线识别方法[J].北京理工大学学报,2014,34(5):485-489.
[8]梁朱冬,陈洪洋.基于改进最小二乘法车道线模型的车道标识线检测识别[J].汽车实用技术,2017(15):98-100.
[9]刘向阳,周淼,郭腾峰.公路行车视距分析与验证[J].中国公路学报,2010,23(S1):36-41.
[10]刘永涛.基于环境感知技术的客运车辆危险行驶状态辨识技术研究[D].西安:长安大学,2015.
[11]傅思勇,吴禄慎,陈华伟,等.综合多畸变因素的摄像机标定[J].仪器仪表学报,2018,39(2):248-256.
[12]陈雨人,付云天,汪凡.基于支持向量回归的视距计算模型建立和应用[J].中国公路学报,2018,31(4):105-113.
[13]李璐璐,赵文川,伍凡,等.摄像机标定中的特征点提取算法研究与改进[J].光学学报,2014,34(5):179-186.
[14]董斌,刘谦.基于三维模型的高速公路中央分隔带外侧车道视距计算方法研究[J].交通科技,2017(1):148-150.
[15]文涛,左东广,李忠科,等.一种鲁棒的高精度摄像机标定方法[J].计算机应用研究,2015,32(11):3489-3491,3495.
[16] ZHANG Z. A flexible new technique for camera calibration[J]. IEEE Transactions on pattern analysis and machine intelligence,2000,22:110-119.