For more information see: https://hforsten.com/third-version-of-homemade-6-ghz-fmcw-radar.html

Compile Tensorflow custom operations in pc/sar/sar_op. See BUILDING file in that directory for the instructions.

pc/sar/process_sweeps.py is used to slice the correct time from the measurement file.

./process_sweeps.py <log file> <start time> <end time> [decimation factor]

It writes sweeps.p file with the data in the format the image formation scripts know how to read. To form the SAR image run ./backprojection_tf.py sweeps.p or ./omegak_tf.py sweeps.p. Some variables related to the image formation can be found on the top of the image formation scripts.

Example data can be downloaded from: https://hforsten.com/fmcw3/parking_lot_sar.log

Use start and end times of 20 and 136 for it. Decimation factor should be 1 or 2 for backprojection and 2 for omega-k.