This is an exploratory project to apply image processing techniques in order to extract data from fisheries observer workbooks. While this project runs on a desktop/laptop computer connected to the internet, the intent is to create something that can be used in an environment where compute and/or bandwidth are highly constrained.
In order to use this, you'll need to find/create an original workbook page as well as an example image to register against it. SPC maintains a library of current master forms within the OFP area of their website. Example "scans" are easier to create using a digital camera or a consumer grade scanner.
An example of using the utility is shown below:
python workbook_reader/align_images.py --scan fake_scan_1.jpg --original ps2_data_page.png --debugExample scans were created by splitting the PDF via MacOS preview. PNG versions of the forms were created with ImageMagick convert:
# density 300 create a huge image that isn't much clearer than 150
convert -quality 100 -density 150 ps2_data_page.pdf ps2_data_page.png
This research is based on examples from the Pyimagesearch Blog.