Closed bryanibit closed 7 years ago
I used Berlin dataset, and then I got the result using the following words.
import matplotlib.pyplot as plt nx.draw(tracks_graph) plt.show()
I still don't understand that due to so many covering objects.
I don't know why the feature.x/y could be negative. Wish I could have an answer. Thank you.
The feature points are in normalized image coordinates, meaning the range is (-1,1).
The origin is located in the center of the photo. Thank you for your answer.
Yes, the position of feature points is stored in normalized image coordinates. I've just added the definition of those in the documentation WIP http://opensfm.readthedocs.io/en/latest/geometry.html#coordinate-systems
@paulinus The doc needs to be finished as soon as possible, which is very useful for the new comers to understand the detail of the project, I think. Especially, we need to add explanations about create_track, reconstruction and mesh, the three python file.
Hello! There is a file called track.csv produced by create_track.py. I want to know the content about that. According to the code
tracks_graph.add_edge(image, str(track_id), feature=(x,y), feature_id=featureid, feature_color=(float(r),float(g),float(b)))
, I think the meaning of track.csv (the following line)DSC07698.JPG 35782 4600 -0.470603 -0.127038 68 84 48 image name track_id feature.x feature.y color.x/y/z
I don't know why the feature.x/y could be negative. Wish I could have an answer. Thank you.