Problem with code

[hello299]

Hi,I don't know the purpose of this code segment,i know u_k and v_k is the image coordinate of the points,but i don't know what is the start_idx for and the why the offset is the proj_idx +1 。
Could you give me some advice? thank you!

[patripfr]

Hi,
In this part of the code, I create the projection map of the undistorted point cloud (see section III G in the paper).
For the undistorted point cloud, multiple points can project into the same pixel. Therefore, we first store all of them, such that we can later select the one with the shortest range, which happens here.

For each pixel we store a maximum of DUPLICATE_POINTS-1 points, such that we can use a predefined size for vector frame.proj_idx .
The structure of this vector is as follows: [offset, i_1, i_2, ... , i_9], where offset stores how many points were already projected in this pixel, and i_i stores the index of the i-th point that was projected in this pixel.
vectorIndexFromRowCol(v_k, u_k) gives the index in frame.proj_idx where the entries for pixel (v_k, u_k) start.
size_t offset = frame.proj_idx[start_idx] + 1; is because we have to increment the start_idx by one since we are adding a point.