For my PhD-research, I have to collect data on bathymetry from the Gulf of Guayaquil. I have found jpg's of maps which show a pattern of numbers representing the bathymetry at certain locations.

Example zoom of one of the maps.

As the region is enormous, I am looking for a way to automatically detect these numbers and read them through Machine Learning (why not take this opportunity to learn something new?). The quality is relatively low and there are many watermarks, lines, different background colors, ...

I have read a lot about the Tesseract OCR but also found examples of using Tensorflow to detect number on images. Could one of those be a valid option? It's no problem to first train a detector. I am quite familiar with the Python language, so I would preferably use a library with a Python API.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.