Distortion coefficient

[ADAS-Iv]

Hello,

I'm trying to use GUI.py but I'm having some troubles with distortion coefficient: I see variation in the chessboard pattern only when I put the coefficients at their max value

[lxzrv]

Hi! Faced with that problem too.
To fix it you should assign dtype of KpCoeff in vcam lib to any np.float type

Hello,

I'm trying to use GUI.py but I'm having some troubles with distortion coefficient: I see variation in the chessboard pattern only when I put the coefficients at their max value

[vgrllc]

I just changed:

self.KpCoeff = np.array([0,0,0,0,0,0,0,0]

to:

self.KpCoeff = np.array([0,0,0,0,0,0,0,0], dtype=float)

in the vcam.py source and it is working.

[b-hakim]

I just changed:

self.KpCoeff = np.array([0,0,0,0,0,0,0,0]

to:

self.KpCoeff = np.array([0,0,0,0,0,0,0,0], dtype=float)

in the vcam.py source and it is working.

Works for me!

Author need to make that change @kaustubh-sadekar

[truher]

+1 where's the vcam source?

[ahtsan]

+1 where's the vcam source?

You can do the following in terminal

python3
>>> import vcam
>>> import os
>>> os.path.dirname(vcam.__file__)

[truher]

indeed. i meant "where is the vcam source in github so i can fix it for everyone?"

[henrikmidtiby]

I have managed to locate the source for the vcam python module and then added it to this fork of the project.

henrikmidtiby@918443d

[ArnobTurja2002Ghosh]

I just changed:

self.KpCoeff = np.array([0,0,0,0,0,0,0,0]

to:

self.KpCoeff = np.array([0,0,0,0,0,0,0,0], dtype=float)

in the vcam.py source and it is working.

You sure it's working accurately just like that? I doubt because the distortion is dependent on the Transformation Matrix calculated in the function update_M() and I think there is a huge mistake in the following block of code:

Rx = np.array([[1, 0, 0], [0, math.cos(self.alpha), -math.sin(self.alpha)], [0, math.sin(self.alpha), math.cos(self.alpha)]])
Ry = np.array([[math.cos(self.beta), 0, -math.sin(self.beta)], [0, 1, 0], [math.sin(self.beta), 0, math.cos(self.beta)]])
Rz = np.array([[math.cos(self.gamma), -math.sin(self.gamma), 0], [math.sin(self.gamma), math.cos(self.gamma), 0], [0, 0, 1]])
self.R = np.matmul(Rx, np.matmul(Ry, Rz))

On what basis did you say that it is working? Is it simply because you see some distortion as you change the distortion coefficients through the trackbar? Have you verified if it is the correct distortion given your input distortion coefficients?