I am trying to georeference a number of PDF documents automatically. I am able to extract the vector data from the PDF - the data is a series of lines representing cadastral boundaries. I can load these vectors into QGIS and it is sitting in screen coordiantes from the PDF. IE: an example (40,50).
I need to find the matching real-world coordinates of these cadastral boundaries, and then apply a transformation to the whole vector dataset generated from the PDF extract.
I am doing this programmatically using Python.
I am trying to identify the commom points between the PDF and the real-world dataset. To do this I have tried the follwoing:
I split the lines in both datasets by vertex to generate lines that are orientated similarily and broken at the same location. I tried to calculate the ratio between orientation and length. This obviosuly did not work as we are dealing with 3 different units (pixel length, measured real world length and degrees). I abandoned this.
I calculated the ratio of the length of the line vs the bounding box area around the line segment. I thought this might be an option, although I cannot seem to find a match in ratios between the two datasets.
Are there any techniques to find common points between the two datasets to allow me to transform the coordinates in the PDF to real-world locations?
