Regarding the published features #154
Description
Hello
I am trying to calculate the z-axis coordinates using the publish feature coordinates. It looks like, in image_processor.cpp, you are publishing the features as homogeneous 2D image plane coordinates after correcting the dataset's distortion, right ? If that's the case, was the distortion model not taken into consideration in Eq.(4) of your paper because it doesn't have an effect on your approach ?
The reason I am asking this is because I am using these coordinates to calculate the depth of the corresponding 3D feature point, which oftentimes turns out negative, and I am trying to figure out if that is due to the rectification of the positive/negative radial distortion model, or to some body frame rotations that I didn't notice along the way.
Your response is deeply appreciated !