Question

在计算机视觉中，特别是计算立体声，我们可以轻松编写算法来查找两个摄像机图像中的对应点。该算法可以用伪代码编写，如下所示：

Repeat for each feature point in the left image {
   calculate the epipolar line in the right image 
   if the epipolar line intersects only one feature point 
   then match those points and remove them from the lists
}
Until no feature point can be matched uniquely

我的问题是如果使用三个摄像头而不是标准的两个摄像头设置，如何更改此算法？

只是一些好的想法或这个伪代码的一些改变版本将是辉煌的。

谢谢。

Answer 1

在任一图像之间有一对匹配的特征点后，您可以确定剩余图像中这些极线的交叉点，并确定最后一个特征。因此，您可以为“第一和第三”以及“第二和第三”相机对重复您的伪代码：

Repeat for each feature point in the first image {
   calculate the epipolar line in the second image 
   calculate the epipolar line in the third image
   if the epipolar line in either image intersects only one feature point {
     calculate epipolar line for matching feature point in the other image. 
     Intersection with epipolar line from first image gives the third point.
     remove triplet from the list
}
Until no feature point can be matched uniquely.

然后

Repeat for each feature point in the second image {
   calculate the epipolar line in the third image
   if the epipolar line intersects only one feature point {
     calculate epipolar line for matching feature point in the first image. 
     Intersection with epipolar line from second image gives the third point.
     remove triplet from the list
}
Until no feature point can be matched uniquely starting from the second image.

在搜索算法中使用极线来查找两个摄像机图像中的对应点

1 个答案: