Overview¶

lcopt-cv: Create fully functional LCA models from hand drawn pictures of process diagrams¶

Lcopt-cv is a python module for creating LCA foreground models from hand drawn pictures of process flow diagrams developed by James Joyce.

Pretty much every LCA starts with drawing a process flow diagram. The difficult bit is turning that diagram into an LCA model which can be analysed.

What if you could just take a picture of the diagram you’ve just drawn and have it instantly turned into an LCA model?

Well now you can - introducing lcopt-cv, computer vision for LCA.

Uses computer vision to generate an LCA model from a photograph of a process flow diagram
Exports model directly to lcopt, allowing models to be analysed using Brightway
Links directly to the ecoinvent or FORWAST databases

Note

Note - lcopt-cv requires the lcopt and brightway2 packages to be installed, and for lcopt to be set up with ecoinvent 3.3 cutoff

The best way to install lcopt-cv is to use the conda package. The command is:

conda install -y -q -c conda-forge -c cmutel -c haasad -c pjamesjoyce lcopt-cv

One additional dependency isn’t available as a conda package and needs to be installed separately using pip. Here is the command:

pip install opencv-python

If you already had lcopt installed and set up - that’s it. If not you need to set up lcopt to talk to brightway.

Full instructions on how to do this are in the lcopt documentation

The short version is

Download the file called ecoinvent 3.3_cutoff_ecoSpold02.7z from the ecoinvent website
Unzip the file using 7zip and make a note of the path of the datasets folder

Run the following command:

lcopt-bw2-setup path/to/ecospold/files # use "" if there are spaces in your path

This will generate the lcopt template databases in brightway2 so that you can analyse your LCA models.

To launch lcopt-cv at the command line type:

lcopt-cv

This will launch the lcopt-cv GUI.

More detailed documentation is on its way… watch this space!