Chenyu
Sure, you can set this option. Note that the constructor of the [DistributedMapperController](https://github.com/AIBluefisher/DAGSfM/blob/dev/src/controllers/distributed_mapper_controller.h), `IncrementalMapperOptions& mapper_options` is passed as a parameter. However, after checking the [code](https://github.com/AIBluefisher/DAGSfM/blob/dev/src/exe/colmap.cc) at line 828-891, minor modification...
Seems very interesting. I may suggest two methods: (1) as one camera is fixed all the time, we can deem the observations (u_k0, u_k1, u_k2, ...) w.r.t. different frames of...
Sure, I think you can do both of the two ways. Since the reference frames are easy to identify by the camera id, both methods should be easy to implement....
It seems your command is not correctly set. What's your `$DATASET_PATH`? It should be your absolute image path. Could you give your running command? So I can help you as...
Seems the database is not read correctly. Did you use the distributed feature extraction and matching method?
> I just created an empty database using colmap's GUI.I browsed your code in 'distributed_mapper_controller.cpp'.I guess it terminated here. >  Try to run [extract_feature.sh](https://github.com/AIBluefisher/DAGSfM/blob/dev/scripts/shell/extract_feature.sh) at first, then start `distributed_sfm.sh`....
> And if I use colmap's GUI to create a database after processing feature extracting and matching, I got some other errors: >  As I modified the database format,...
I would suggest to set logs to locate where the bug occurs, so that I can help as I could.
I mean to set log information in the code to debug where the dump occurs, or it would be a time for me to reproduce this bug.
> I used 50 images per cluster with 100 images. it group to into clusters of 66 images, 2 clusters. But the clusters only ran on 1 worker not both...