-
Notifications
You must be signed in to change notification settings - Fork 1.7k
Add ObservationManager #2575
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add ObservationManager #2575
Conversation
|
||
const double tri_angle = CalculateTriangulationAngle( | ||
proj_center1, proj_center2, point3D.xyz); | ||
EXPECT_GE(tri_angle, DegToRad(kMinTriAngleDeg)); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This test sometimes fails on mac and Windows with the previous value of 1°. It seems that the sampling of camera poses doesn't constrain them to have a sufficient angle.
Looks great to me. I assume most of the code in the ObservationManager is untouched and just moved from Reconstruction/Image classes? Have you run through a dataset and checked that it produces the exact same number of final observations? |
PTAL. I have indeed mostly moved around code without change of logic. I have checked that the result is identical in terms of reconstruction statistics. |
Perfect, thanks a lot. |
This removes the filtering logic in
Reconstruction
as well as the SfM-specific data inImage
, such as the visibility pyramid. The logic is unchanged except one minor point: I removed thenum_observations
column in the GUI database image viewer because this field is now only available during reconstruction. It is anyway mostly redundant withnum_points2D
.