|
|||||
|
|
| 交通事故摄影图像的自标定三维重建及其实现方法 | |
| 引用本文： | 鲁光泉,许洪国,马社强,李一兵.交通事故摄影图像的自标定三维重建及其实现方法[J].公路交通科技,2006,23(9):108-112. |
| 作者姓名： | 鲁光泉 许洪国 马社强 李一兵 |
| 作者单位： | 1. 清华大学,汽车安全与节能国家重点实验室,北京,100084 2. 吉林大学,交通学院,吉林,长春,130025 3. 中国人民公安大学,交通管理工程系,北京,102614 |
| 摘 要： | 为避免测量过程中在现场设置标定参考点,研究了以基础矩阵估计为基础的交通事故摄影图像自标定模型,并分析了其实现过程。把便于递推实现的无限对称指数滤波器引入平均梯度矩阵计算,实现了图像特征点的检测;通过匹配强度系数计算,对两幅图像的特征点进行了匹配;分析了基础矩阵计算及不良匹配点剔除的最小中值二乘法;对通过本质矩阵分解标定摄影图像的实现过程进行了论述。通过实验对模型及其实现方法进行了验证。 |
| 关 键 词： | 交通安全 交通事故 图像分析 三维重建 自标定 基础矩阵 |
| 文章编号： | 1002-0268(2006)09-0108-05 |
| 收稿时间： | 2005-05-17 |
| 修稿时间： | 2005年5月17日 |
Three-dimension Reconstruction of Traffic Accident Photographs by Self-calibration |
|
| LU Guang-quan,XU Hong-guo,MA She-qiang,LI Yi-bing.Three-dimension Reconstruction of Traffic Accident Photographs by Self-calibration[J].Journal of Highway and Transportation Research and Development,2006,23(9):108-112. | |
| Authors: | LU Guang-quan XU Hong-guo MA She-qiang LI Yi-bing |
| Abstract: | To avoid sitting reference object on scene,the model of self-calibration based on fundamental matrix estimation and its implementation are proposed.Feature points detection and feature points matching were studied.The Infinite Symmetric Exponential Filter(ISEF) was used as 2D weight matrix of average gradient matrixes to detect feature points.A formula of match strength was used to match feature points.The least median of squares(LMedS) was used to estimate fundamental matrix.The process of essential matrix decomposing to achieve project matrixes for three-dimension reconstruction was discussed.The test results verified the method. |
| Keywords: | traffic safety traffic accident analysis of photographs three-dimension reconstruction self-calibration fundamental matrix |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |