论文部分内容阅读
由于交通事故现场测量对象呈静态特性 ,根据计算机视觉中的双目视觉原理 ,可以使用两幅从不同方位拍摄的事故现场图像对事故现场进行测量。在实现过程中用预制三维方块标定物对拍摄使用的照相机内外参数进行标定 ,求取投影变换矩阵 ,并用基于面积的配准方法进行对应点匹配。模拟事故现场的测量结果表明用该方法对交通事故现场进行测量 ,可为事故再现提供大量数据
Due to the static nature of the measured objects at the scene of an accident, two accident scene images taken from different directions can be used to measure the scene of the accident according to the binocular vision principle in computer vision. In the process of implementation, the pre-fabricated three-dimensional block calibrator is used to calibrate the camera’s internal and external parameters used in shooting. The projection transformation matrix is obtained and the corresponding point matching is performed by the area-based registration method. The measurement results of the simulated accident scene show that using this method to measure the scene of an accident can provide a large amount of data for the accident reproduction