izhengfan / se2lam

(ICRA 2019) Visual-Odometric On-SE(2) Localization and Mapping
https://github.com/izhengfan/se2lam
MIT License
404 stars 108 forks source link

Why does Jacobian calculation only use fx and not fy in optimization #30

Closed CrisGao closed 4 years ago

CrisGao commented 4 years ago

I found that the Jacobian calculations in the program only involve fx, such as SE(2)-XYZ constrained Jacobian calculations: src/EdgeSE2XYZ.cpp/linearizeOplus() image and Thirdparty/g2o/g2o/types/sba/types_six_dof_expmap.cpp/linearizeOplus() image

izhengfan commented 4 years ago

We use only fx because in the early-stage we use the types of g2o, and g2o::CameraParameters only has one parameter focal_length, i.e. they consider fx==fy. This is correct in all the sensors we met with, so we just keep consistent to it.

CrisGao commented 4 years ago

I got it,Thank you