Skip to content

Refining Self-Calibration with COLMAP BA Module #3196

@fengYangwcrcsvhw

Description

@fengYangwcrcsvhw

Dear COLMAP authors,

Hello!

I am currently attempting to use the BA (Bundle Adjustment) module in the COLMAP library to perform a final refinement on self-calibration results. Specifically, I plan to provide the following initial data:

Camera intrinsics and distortion parameters initial estimates (strictly adhere) Intrinsics: [fx, fy, cx, cy] Distortion parameters: [k1, k2, p1, p2] (in my application, the camera intrinsics and distortion parameters are identical for every frame, so there is only one global set of initial parameters)

Initial estimates of external parameters for each image For the general case, [R, T] is provided; If the camera is undergoing pure rotation, then only [R] is provided

Correspondence data For each frame, the correspondence between 3D points (X, Y, Z) in the global coordinate system and 2D points (x, y) in the image coordinate system is provided

My goal is to obtain through BA optimization: Optimized camera intrinsics and distortion parameters (since they are the same for every frame, only a single set is required) (strictly adhere); Optimized external parameters for each frame ([R, T] or [R] for pure rotation); As well as the overall reprojection error metric.

Could you please clarify if I can directly use the existing BA module in COLMAP to achieve such a refinement process? If yes, is there already an appropriate interface provided, or would custom modifications be necessary? I would appreciate any advice or considerations regarding this usage. Also, is there already example code available?

Thank you very much for your time and help!

Sincerely,

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions