This paper investigates a new method of reconstructing moving track from the measured value over several views when the object moves along line in 3D space. Here the camera can be one moving camera

or a set of camera being located at different position. This problem was put forward by A.Shashua first and was defined as “trajectory triangulation”. Under some restriction of the track of moving

they determined the line in 3D (moving track) from the relativity of point and line and by the aids of Grassmann-Carley algebra and Plücker coordinate. This paper brings forward a new method based on the rank restriction of the "measured matrix"

that is to say if the points are on a line

the matrix formed by the coordinates of these points must have rank 3. In order to simplify the computation

the camera matrices have been transformed so that the last column of every projective matrices has only one no zero value. The algorism of the method is evaluated on both synthetic and real world images. Comparing with the method in[1]

our method is simply and directly

and it can be realized in a general projective coordinate system. Moreover

it can be generalized to any situation as long as the shape of trajectory of moving object is polynomial

although only moving along line is investigated in this paper.