Bundle Adjustment for Matlab

This code package constaints Bundle Adjustment software with support for multiple residual terms. If you use this software please cite [1].

The internal camera calibration can be set fixed or as part of the optimization. Multiple sets of views with shared calibration can be defined. 3D points and cameras can (independently) be set as fixed or part of the optimization.

Every point can be assigned to [0, inf] planes. The 3D point-plane projection error is used as a optimization residual. The plane normals can be set fixed or as part of the optimization.

Same as 2) but for camera location. This is useful if a camera plane / ground plane is known. This is used in: [1].

Use mutual camera-camera visibility as 3D (camera location) to 2D (observed camera location) reprojection constraint. Otherwise similar to 1). Ref: [2].

For given line segments in image, and associated VPs (3D direction unit vectors) use VP reprojection in image as residual. Every VP can be set fixed or as part of the optimization. Ref: [3].

A smoother matrix (Reinsch Form, See [4]) is used to penalize non-smooth camera paths. This is used in: [1].

1. T. Kroeger, L. Van Gool, Video Registration to SfM Models, ECCV 2014.
2. T. Thormählen, et al., Exploiting Mutual Camera Visibility in Multi-Camera Motion Estimation, ISVC 2009.
3. J.-P. Tardif, Non-Iterative Approach for Fast and Accurate Vanishing Point Detection, ICCV 2009. Fig. 4.
4. T. Hastie, et al. The Elements of Statistical Learning, Data Mining, Inference, and Prediction. Springer, 2nd, 2009 edn. (2009), Page 154 f.

1. Download and install the Ceres Solver: http://ceres-solver.org/

Make sure to activate shared library support. Last tested with ceres v1.8.0.rc2.

2. Adapt: src/CMakeLists.txt and src/mex_link.sh.

Change $MATLAB_INCLUDE_DIR in CMakeLists.txt to you local Matlab include directory. Change $CERESLIBSPATH in CMakeLists.txt to you Ceres install directory if it is not already visible from matlab. In mex_link.sh change path to your local matlab binary.

3. Compile with

cmake && make

Test binary BAdjustbin with make test. Test mex-binary BAdjustMex by opening MATLAB and running BATestSuite/run_BA_TestSuite.m.

4. Currently no detailed documentation is available.

From matlab you can call the library directly with BAdjustMax(data), with 'data' being the input and output structure for a BA problem. See examples in BATestSuite/run_BA_TestSuite.m of how the datastructure is created and filled. Please make sure all data types, dimensions are correct in the input struct. Input verification is still very basic.

You can also call the libray over the shell by ./BAdjustbin inputfile outputfile. Please see BATestSuite/func_readwrite_BA_problem.m for example code for writing and parsing of the text files.

If you need to call the library from C directly, have a look at src/ceresBA_datastruct.h for documentation of the data structure, and src/ceresBAbin.cpp of an example of how the solver is called.

If you use this software please cite [1].

GPLv3: http://gplv3.fsf.org/

All programs in this collection are free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.