何扬名, 杜建强, 汪灿华, 熊光耀. 虚拟双目视差法在公交客流统计中的应用[J]. 微电子学与计算机, 2014, 31(5): 90-93.
引用本文: 何扬名, 杜建强, 汪灿华, 熊光耀. 虚拟双目视差法在公交客流统计中的应用[J]. 微电子学与计算机, 2014, 31(5): 90-93.
HE Yang-ming, DU Jian-qiang, WANG Can-hua, XIONG Guang-yao. Application of Virtual Binocular Disparity in Passenger Flow Statistics on Bus[J]. Microelectronics & Computer, 2014, 31(5): 90-93.
Citation: HE Yang-ming, DU Jian-qiang, WANG Can-hua, XIONG Guang-yao. Application of Virtual Binocular Disparity in Passenger Flow Statistics on Bus[J]. Microelectronics & Computer, 2014, 31(5): 90-93.

虚拟双目视差法在公交客流统计中的应用

Application of Virtual Binocular Disparity in Passenger Flow Statistics on Bus

  • 摘要: 提出了一种检测运动目标的新思路:虚拟双目视差法.用一个固定的摄像机检测运动目标,拍摄连续的两帧图像,可以将这种情形看作:目标是静止的,而摄像机在两个不同的位置对目标进行拍摄.于是连续两帧图像,可以看成是一个摄像机在原来地方拍摄的一帧图像和另一个摄像机在目标运动反方向的目标运动距离处拍摄的图像.虚拟双目视差法就将单摄像机的运动目标检测转化为双摄像机的静态目标检测.利用双目视觉的图像匹配技术对连续两帧图像进行匹配,匹配所求得的双目视差大小就是运动目标的运动距离,双目视差的反方向就是目标的运动方向.该方法通过实际公交车的序列图像进行了检测,证明了其良好效果.

     

    Abstract: A new idea to detect moving targets is put forward,which is virtual binocular disparity algorithm.Normally,one fixed camera is used to detect moving targets,shooting two continuous images.This condition can be considered as the condition in which targets are stationary and two cameras are put in two different positions.So two continuous images can be seen as one shot at the original position and the other shot at the position opposite to the moving target.Virtual binocular disparity algorithm transforms one camera detecting moving targets to two camera detecting stationary targets.Using matching technique of binocular on two continuous images,the disparity can be gotten.The disparity describes the moving features of the moving targets including moving distance and direction.The good effect of this algorithm is proved by practical images on bus.

     

/

返回文章
返回