Bug in interpMapValueWithDerivatives (OccGridMapUtil.h)?

[alik55]

It seems x and y weights are switched in derivatives at Line 341:

return Eigen::Vector3f(
((intensities[0] * xFacInv + intensities[1] * factors[0]) * (yFacInv)) +
((intensities[2] * xFacInv + intensities[3] * factors[0]) * (factors[1])),
-((dx1 * xFacInv) + (dx2 * factors[0])),
-((dy1 * yFacInv) + (dy2 * factors[1]))

should be (changes in the last two lines):

return Eigen::Vector3f(
((intensities[0] * xFacInv + intensities[1] * factors[0]) * (yFacInv)) +
((intensities[2] * xFacInv + intensities[3] * factors[0]) * (factors[1])),
-((dx1 * yFacInv) + (dx2 * factors[1])),
-((dy1 * xFacInv) + (dy2 * factors[0]))

[ytgui]

Hi,
I also find that bug these days, and I'm confused why there is a math error but this lib could run correctly.
I think that because xFacInv or yFacInv are both weight value in range (0,1), and dx,dy decide the positive or negative. A multiplication of dx1*xFacInv is same to k * dy1 * yFacInv.

[watcherZY]

i have confirmed that you are correct , i used a same rosbag to test -((dx1 * xFacInv) + (dx2 * factors[0])),
-((dy1 * yFacInv) + (dy2 * factors[1])) AND -((dx1 * yFacInv) + (dx2 * factors[1])),
-((dy1 * xFacInv) + (dy2 * factors[0])). the second one works much better! and dosen^t become disorder when the robot rotation

[jondo]

Why was this closed?

[StefanFabian]

I've closed all issues older than 5 years due to their age but you're right, this one may still be relevant.
I'm not sure if @skohlbr will have the time to look at this, though.