LeafEye

Leaf thermal image analysis assistant

Allows for filtering of subject data at high resolutions from thermal image datasets. Useful for when you want to acquire thermal distribution and not just point measurements.

Software requirements

Python 3 (to run this)

I recommend using Anaconda to create a Python 3 environment.

Tested on Mac (10.10 and higher) and Linux (Ubuntu)

Recommended but not required:

ImageJ for making image masks

Usage

Setting up files

Export each photo from FLIR Tools in grayscale, and remove extra marks (e.g. point measurements).
Create mask image in ImageJ
Select the area you want with the Polygon tool
Fill
Clear Outside
Selection > Select None (or just click outside the polygon)
Invert (The area of interest should be in black and the areas to ignore in white)
(If your object has a clear background, you can also try using Threshold instead.)
Save As ... > Jpeg
Repeat for each image in dataset to be processed. (The pipeline and further statistics will likely take some time to process, so you might want to make a minimal working dataset to test your full analysis pipeline on)
Encode sample name, image filename, min/max temperature, and mask filename in a table (see template)
Ensure that all the images and input_batch.csv are in the same folder.

Processing images

Run batch.py on the Terminal with python batch.py <folder path>. On Mac, you can insert the folder path by dragging the folder on to the Terminal window.
You should get some messages like this:

Initializing ... 
Starting image 1............done!
Starting image 2............done!
Saving csv ...
Completed!

Look in the folder for an output.csv. This file can now be further processed by your statistical package.

mdchia / LeafEye