Macrobot is an image analysis software for studying plant-pathogen interactions on macroscopic level. Currently the macrobot software can detect and quantify the following plant-pathogen interactions:
- Barley powdery mildew (Blumeria graminis f. sp hordei)
- Wheat powdery mildew (Blumeria graminis f. sp tritici)
- Wheat yellow (stripe) rust (Puccinia striiformis f.sp. tritici)
- Wheat brown (leaf) rust (P. graminis f. sp. tritici)
Figure 1: Powdery mildew on barley plants
The hardware system is based on a custom fully automated multispectral 2D imaging station (Figure 2).
Figure 2: Macrobot Module
See the macrobot hardware in action: https://www.youtube.com/watch?v=SmoKQ_uMp34&t=56s
The entire pipline from image aquisition to image analysis is shown in Figure 3.
Figure 3: Software pipeline
Lueck et al., (2020). BluVision Macro - a software for automated powdery mildew and rust disease quantification on detached leaves.. Journal of Open Source Software, 5(51), 2259, https://doi.org/10.21105/joss.02259
https://macrobot.readthedocs.io/en/latest/index.html
Macrobot software was build and successfully tested on Windows operation system (Windows 7 and 10).
->Install Anaconda (https://www.anaconda.com/distribution/)
conda create --name macrobot python=3.7
conda activate macrobot
conda install pip
pip install macrobot
- Create a folder for the result. We will create a new folder on the desktop called mb_results.
- Open the Ananconda prompt and activate your macrobot environment if you are not already there.
conda activate macrobot - Macrobot is a command line program which requires the following arguments:
- source path (-s) - the path with the images coming from the Macrobot hardware system
- destination path (-d) - the path to store the results
- pathogen (-p) - which pathogen to predict ("mildew" or "rust")
- For a test case we will use a test image set which will be automatically downloaded by the start of the software. To tell the software to use the test images, we will enter "test_images" for the source path -s argument
- Start the software with the following command for mildew (adapt the destination path):
mb -s test_images -d C:\Users\name\Desktop\mb_results\ -p mildew
or rustmb -s test_images -d C:\Users\name\Desktop\mb_results\ -p rust - In your destination folder should appear all results:
- A csv file with the predicted values per leaf
- A report html file in folder report which allows and easy control over the pipeline.
- Images created by the software (white=pathogen, red=leaf detection, black=background)
If you want to use a real world experiments, make sure to provide the following folder structure with five images per plate (see documentation)
cd to installation path and test folder e.g. d:\Anaconda\envs\mb_test\Lib\site-packages\macrobot\tests
Run pytest:
pytest
We are strongly looking for contributions, some ideas how to support our software could be found here: https://github.com/snowformatics/macrobot/wiki/Contributions