uzh-rpg/rpg_ig_active_reconstruction

Question on Visibility Likelihood

YiduoWang47 opened this issue · 1 comments

To whom that may concern,

Hello!
I am working on an active mapping system that utilises information gain, and have implemented the formulations proposed in this work as the state-of-the-art.
Going through the code base, I have noticed that the visibility likelihood (p_vis_), such as line 93 in occlusion_aware.inl, is computed as the product of occupancy probability (p_occ). In the paper, on the other hand, p_vis_ is defined as the product of the (1 - p_occ). I wonder whether I misunderstood the code.
Any and all help will be much appreciated!

@YiduoWang47 Hi, I noticed the same problem, do you figure out why is this happen?