How to reduce turn-around time for development for Sequoia camera

In order to make the OpenDroneMap compatible with multiband observations using the Sequoia camera, I have made some modifications to the source code.

After making some additions and modifications to the source code such as “multispectral.py” and “photo.py” in the local environment, I built it with the following command.

docker build -t my_odm_image .

Even if I modify just one line, it takes a long time to rebuild everything, including downloading the source code. (MacBookPro, docker, about 25 minutes)

I’d like to reduce the turn-around time for development. Does anyone know how I can build only the parts I need and create a docker image?

Even if it only takes 20 minutes instead of 30 minutes, it will help a lot.

===
In addition, the OpenDroneMap built in my local environment is almost ready for the Sequoia camera, including the readout of the solar radiation sensor.
It is now possible to get the correct reflectance map values without having to take calibration shots every time.

I will make a pull request when the source code modification is finished.

Also, the current official version of ODM 2.5.7 has some problems with radiometric calculations, and it seems that Micasense and P4Multispectral do not give the correct reflectance and NDVI map values.

1 Like

Awesome work!

I think ccache would help, but I don’t know how to use that within docker. I usually have it on the Linux host.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.

You can use docker cp to copy the python right into the docker image, which will save you a rebuild. If there’s also pyc associated with the file, you’ll need to build that and copy that in.

1 Like

Thanks for your advice. I’m not familiar with python or docker, so I’ll take some time to become proficient.

1 Like