It’s been a while since my last post, but finally there’s something to report! I’ve created a new GitHub repository where I maintain tools for map classification in Python.
Check it out here.
As much as I like working with R, coding in Python has proven a much faster and more convenient alternative for working with image data, primarily because it allows using the excellent OpenCV toolkit. Currently, the new GitHub repository hosts a loosely connected set of Python modules that are useful for automatic map classification. Among them is a PixelClassifier class, which essentially implements a Python version of the supervised classification documented here, as well as some preliminary tools for character detection and recognition. I’m currently working on additional tools for text removal, as well as line and shape detection, which I’ll make available once they work reliably.