The method of calculating the dispersion parallax place in stereovision

摘要

The method of calculating the dispersion parallax place when epi- Pola geometry is known in the system (100) for stereovision which has 2 cameras (T₁, T₂) which photograph the same scene, is executed. As for this method, from the position where two differs depending upon 2 cameras the 1st picture of the particular scene and the step which takes in the 2nd picture (IM₁, IM₂); Being the step which selects at least one pixel in the 1st picture, the aforementioned step which includes one point where the aforementioned pixel relation is attached by one point of the particular scene, as for the 2nd picture this in the aforementioned point of the particular scene relation is attached; And calculating the displacement to the aforementioned point inside the 2nd picture from the aforementioned pixel, it includes the step which minimizes cost function. This cost function includes with the section which depends on the difference with the 1st picture and the 2nd picture and the section which depends on the distance to the aforementioned point inside the 2nd picture from the epi- Pola straight line. The case where the camera is proofread it checks whether or not in order to consider error or uncertainty, it is generic to the approved territory around subset in the epi- Pola straight line where existence of the aforementioned point is allowed. Selective figure Figure 2