Speedy Measurement of Arabidopsis Rosette Traits (SMART)
Author: Suxing Liu
Robust and parameter-free plant image segmentation and trait extraction.
- Process with plant image top view, including whole tray plant image, this tool will segment it into individual images.
- Robust segmentation based on parameter-free color clustering method.
- Extract individual plant gemetrical traits, and write output into excel file.
Requirements
Either Docker or Singularity is required to run this project in a Unix environment.
Usage
Docker
docker pull computationalplantscience/smart
docker run -v "$(pwd)":/opt/arabidopsis-rosette-analysis -w /opt/arabidopsis-rosette-analysis computationalplantscience/arabidopsis-rosette-analysis python3 /opt/arabidopsis-rosette-analysis/trait_extract_parallel.py -i input -o output -ft "jpg,png"
Singularity
singularity exec docker://computationalplantscience/arabidopsis-rosette-analysis python3 trait_extract_parallel.py -i input -o output -ft "jpg,png"
Contents
Requirements
The easiest way to run this project is with Docker or Singularity .
To pull the computationalplantscience/smart
image, the current working directory, and open a shell with Docker:
docker run -it -v $(pwd):/opt/dev -w /opt/dev computationalplantscience/smart bash
Singularity users:
singularity shell docker://computationalplantscience/smart
Usage
Segmentation
To perform color segmentation:
python3 /opt/smart/core/color_seg.py -p /path/to/input/file -r /path/to/output/folder
You can also pass a folder path (-p /path/to/dir
). By default any JPG
and PNG
are included. You can choose filetype explicitly with e.g. -ft jpg
.
To extract traits:
python3 /opt/smart/core/trait_extract_parallel_ori.py -p /path/to/input/file -r /path/to/output/folder
You can also use a folder path as above, likewise for filetype specification.
By default this script will not perform leaf segmentation and analysis. To enable leaf analysis, use the -l
flag.
To indicate that your input is a multiple-tray or -individual photo, add the -m
flag.