distort image or undistort image?

[Shuyan]

Hello,

I am currently use this toolbox and dig into the code a little bit. I have a question regarding to the intrinsic parameters.

Based on the code, the LED/Laser light is detected from images. Are these images directly from cameras or undistorted with intrinsic parameters in advance?

I know the toolbox is designed to calculate intrinsic and extrinsic parameters at the same time. Therefore, there are several points I can not understand and hope to discuss with you here.

1.       1. When you “%estimate the fundamental matrices and epipoles with the method of [Har95]”, will the corresponding points from diffrernt cameras follow stereo geomantic?  Because if they are extracted from original (distorted) images, they are not perspective images. Can one simply extract Fundamental matrix and epipolar out of them?

2.       2. If I already calculated the intrinsic parameters for several/all cameras individually, would it possible to use these values as initial parameters? Since the toolbox optimize intrinsic and extrinsic at same time, too many variables will easily stuck the optimization operation at local minimum. Providing an initial value can help to reduce this problem. If this is possible, could you please tell me which function, where I should modify in the toolbox?

Regards,

Shuyan Liu