Command Line Reference

The kelp-o-matic package includes one command line tool, kom. It will be registered in the same Conda environment that the kelp-o-matic package is installed to.

$ kom --help

 Usage: python -m kelp_o_matic.cli [OPTIONS] COMMAND [ARGS]...

   Options
  --version             -v
  --install-completion          [bash|zsh|fish|powershell|pwsh]  Install completion for the specified shell. [default: None]
  --show-completion             [bash|zsh|fish|powershell|pwsh]  Show completion for the specified shell, to copy it or customize the installation. [default: None]
  --help                -h                                       Show this message and exit.

   Commands
  find-kelp           Detect kelp in image at path SOURCE and output the resulting classification raster to file at path DEST.
  find-mussels        Detect mussels in image at path SOURCE and output the resulting classification raster to file at path DEST.

find-kelp

$ kom find-kelp --help

 Usage: python -m kelp_o_matic.cli find-kelp [OPTIONS] SOURCE DEST

 Detect kelp in image at path SOURCE and output the resulting classification raster to file at path DEST.

   Arguments
  *    source      TEXT  Input image with Byte data type. [default: None] [required]
  *    dest        TEXT  File path location to save output to. [default: None] [required]

   Options
  --species        --presence             Segment to species or presence/absence level. [default: presence]
  --crop-size                    INTEGER  The data window size to run through the segmentation model. [default: 1024]
  --rgbi           --rgb                  Use RGB and NIR bands for classification. Assumes RGBI ordering. [default: rgb]
               -b                INTEGER  GDAL-style band re-ordering flag. Defaults to RGB or RGBI order. To e.g., reorder a BGRI image at runtime, pass flags `-b 3 -b 2 -b 1 -b 4`. [default: None]
  --gpu            --no-gpu               Enable or disable GPU, if available. [default: gpu]
  --tta            --no-tta               Use test time augmentation to improve accuracy at the cost of processing time. [default: no-tta]
  --help       -h                         Show this message and exit.

Example

Classify kelp species in an RGB image:

kom find-kelp --species --crop-size=1024 ./some/image_with_kelp.tif ./some/place_to_write_output.tif

Classify kelp presence/absence in an BGRI image:

kom find-kelp --rgbi -b 3 -b 2 -b 1 -b 4 --crop-size=1024 ./some/image_with_kelp.tif ./some/place_to_write_output.tif

Tip: Reduce windowed processing artifacts

To reduce artifacts caused by Kelp-O-Matic's moving window classification, use the largest crop_size that you can. Try starting with a crop_size around 3200 pixels and reduce it if your computer is unable to load that much data at once and the application crashes.

We hope to improve Kelp-O-Matic in the future such that the maximum crop size for your computer can be determined automatically.

CLI Example:

kom --crop-size=3200 your-input-image.tif output-image.tif

Also worth noting is that the crop_size parameter will change the outputs. If you need to reproduce a result, make sure to use the same crop_size as the original run.

find-mussels

$ kom find-mussels --help

 Usage: python -m kelp_o_matic.cli find-mussels [OPTIONS] SOURCE DEST

 Detect mussels in image at path SOURCE and output the resulting classification raster to file at path DEST.

   Arguments
  *    source      TEXT  Input image with Byte data type. [default: None] [required]
  *    dest        TEXT  File path location to save output to. [default: None] [required]

   Options
  --crop-size                  INTEGER  The data window size to run through the segmentation model. [default: 1024]
               -b              INTEGER  GDAL-style band re-ordering flag. Defaults to RGB order. To e.g., reorder a BGR image at runtime, pass flags `-b 3 -b 2 -b 1`. [default: None]
  --gpu            --no-gpu             Enable or disable GPU, if available. [default: gpu]
  --tta            --no-tta             Use test time augmentation to improve accuracy at the cost of processing time. [default: no-tta]
  --help       -h                       Show this message and exit.

Example

kom find-mussels --crop-size=1024 ./some/image_with_mussels.tif ./some/place_to_write_output.tif