Given point-cloud data : For example. position + color and normal.
What will you use to generate mesh of this data and keep reasonable quality for real time cases ?
Additional info to consider:
In some cases given also 3D cameras world transforms which generated the PCL.
Number of input vertices : ~50k and up to millions .
Target latency from input to output : < 40ms
Output triangle mesh (preferably not a must)
More info about mesh reconstruction
A Benchmark for Surface Reconstruction
Any inputs are welcome.
How the data look like ?
- Float* Position [x1,y1,z1,x2,y2,z2...]
- Float* Normal [nx1,ny1,nz1,nx2,ny2,nz2...]
- float or int8 for the color.