Openmvg: Help with error: "There is no defined intrinsic data in order to compute an essential matrix for the initial pair."

Created on 3 Jun 2018  路  3Comments  路  Source: openMVG/openMVG

When running the sequential pipeline with certain image sets, the pipeline stops at the stage linked below with the error

There is no defined intrinsic data in order to compute an essential matrix for the initial pair.

https://github.com/openMVG/openMVG/blob/db92617fea420363abcbad1cb3773706f75012ad/src/openMVG/sfm/pipelines/sequential/sequential_SfM.cpp#L226

Although I get an idea of why, could someone help me understand why that is happening and how to prepare my data set to avoid this? Thanks.

Most helpful comment

Your camera model is not registered in the camera database.
If all your images have the same size, you can specify an approximate focal length to SfM_InitImageListing by using the -f option.
The value to use as an argument can be 1.2 * max(image_width, image_height)

All 3 comments

Your camera model is not registered in the camera database.
If all your images have the same size, you can specify an approximate focal length to SfM_InitImageListing by using the -f option.
The value to use as an argument can be 1.2 * max(image_width, image_height)

when i can change your option in the code tutorial_demo.py or when ?
Your camera model is not registered in the camera database.
If all your images have the same size, you can specify an approximate focal length to SfM_InitImageListing by using the -f option.
The value to use as an argument can be 1.2 * max(image_width, image_height) please help me

You can modify the python file here: https://github.com/openMVG/openMVG/blob/develop/src/software/SfM/tutorial_demo.py.in#L49
default is "-o", matches_dir, "-d", camera_file_params, "-c", "3"])
You can add your new options at the end
"-o", matches_dir, "-d", camera_file_params, "-c", "3", "-f", "<YOUR NUMBER>"])

Was this page helpful?
0 / 5 - 0 ratings

Related issues

itsdsk picture itsdsk  路  6Comments

learnmano picture learnmano  路  6Comments

rttgnck picture rttgnck  路  6Comments

autosquid picture autosquid  路  4Comments

kalosma picture kalosma  路  4Comments