Python 3.5+ is the only supported Python version and there are several modules required for this to run. The GUI is built with tkinter but the main functionality comes from openCV. You will will need to build openCV from source yourself, including the contrib modules found at https://github.com/opencv/opencv_contrib. Find a tutoiral online on how to install. The remaining packages can be seen in the stitcher.py and mosaic.py files.
The overlap is not perfect everytime. To minimise errors, try to take images with at least 1/3 overlap in both directions. Even if you do your best, sometimes the placement is incorrect. There is currently no Undo function, so if you get a misplaced tile, unfortunately you will need to start again. However, this happens < 1% of the time and only where overlapping regions are very similar. If you want perfection, feel free to play around under the hood.
In a terminal run python3 mosaic.py. Images from a microscope should be saved in the Input folder of this parent directory.
Pressing ⌘ + G will begin checking the Input folder of images and stich them based on their sequential order.
Note: Images should be saved using a numerical convention, and the number of files in the Input folder should not exceed 9 at anyone time. This is to avoid files Image_1.png and Image_10 being grouped sequentially. Ideally you want to have no more than 1 to 2 images in the Input folder at anytime if you are imaging and stitching simultaneously.
Images will be read in and stitched to the canvas. A green box shows where the latest image is in the context of the full mosaic.
Pressing ⌘ + G will stop looking for images to add to the mosaic. Mosaics can be saved using the Save button. An error will occur if a correct filename is not entered. i.e anything other than *.png, *.tiff or *.jpg. Mosaics are saved in the Output folder.
Pressing ⌘ + R will reset the canvas to blank.
A resized full-sized stitch can be seen here
