Philipp Glira
Philipp Glira
simpleICP is (currently) for the fine alignment of the point clouds only, i.e. the point clouds must be roughly aligned in its initial state. If that's not the case, the...
Hi, I am highly interested in porting simpleICP to PyTorch. However, I'm not very familiar with PyTorch so I need to start from the very basics first.
Same issue here. There seems also to be a large time delay between the sensor timestamps, e.g. radar and lidar.
One of the issues is that leading zeros have been omitted in the timestamps file. An Example from ``Navtech_Cartesian.txt``. **Original**: Frame: 000001 Time: 1549880486.86608439
Well yes, but only by checking the numbers of digits after the comma and adding a leading zero if necessary. However, generally, I lost interest in this dataset as it...
Interesting. Are both planes at z=0? Maybe simpleICP simply fits two identical planes at z=0 and consequently all residuals are equal to 0. Can you share point clouds and code?
Please keep in mind that simpleICP matches planes, not points. Consequently, rather dense point clouds are needed to estimate the planes from a certain number of points (parameter ``neighbors``). Thus,...
Yes, you're right. Do you have the time for a PR?
Yes, of course, with this I can try to fix it. Thanks.
I'm also having the same issue with an ouster lidar. Any suggestions or workarounds?