search

PRBonn / semantic-kitti-api

SemanticKITTI API for visualizing dataset, processing data, and evaluating results.
http://semantic-kitti.org
MIT License
783 stars 187 forks source link

Evaluate by Distance Program #62

Closed tsunghan-wu closed 4 years ago

tsunghan-wu commented 4 years ago

To the best of my knowledge, data of each scan contains (x, y, z, intensity) four dimention information. Thus, I think only the first three dimension is needed when calculating depth. I would like to know if the code here is correct? Thanks!

jbehley commented 4 years ago

Good catch. You are correct. I think it should be:

depth = np.linalg.norm(scan[:, :3], 2, axis=1) # get depth to filter by distance

However, the intensity is as far as I remember in [0,1] and therefore I would not expect a big difference between the results. I still will change it for consistency.

tsunghan-wu commented 4 years ago

I think you're right. I've added it on and created a pr!

jbehley commented 4 years ago

I just merged it, thanks.