Code for the porejct-seminar "Malen nach Zahlen mit dem X-Y-Plotter (Teil Bildverarbeitung)" at TU Darmstadt (SS-19). The code in this repository is organized as follows:
nbs: Jupyter notebooks that mainly contain initial experiments and some parameter tuning widgets (e.g.circle_detection_acc.ipynb).data: Test images, csv files and stored debug output.scripts: Contains the main program (run_detection_pipeline.py) and code that produces evaluation results and plots. For more information see the documentation in the source files.
We require a working installation of OpenCV (3.x or 4.x), Tesseract (Version 4.0, including pytesseract), Numpy and Matplotlib (Supplies GUI components). It is recommended that you install python packages into a virtual environment.
To run the detection pipeline execute the script run_detection_pipeline.py in your preferred terminal. The script will output a header-less csv file containing the ordered list of detected coordinates and line types.
You need to specify the following parameters:
--image: The file path to the input camera image.--outFile: A file path where the output csv file should be written to.--debug(optional): A flag that when present will lead to the script also writing debug images to the same directory as--outFile.