Aljoscha Leonhardt (leonhardt@neuro.mpg.de, Borst Lab, MPI of Neurobiology)

This is code for a 3D fly tracking arena that allows simultaneous monitoring of multiple Drosophila as well as real-time display of 3D stimuli. Note that this software is not optimized to be usable on arbitrary set-ups and unfortunately we currently don't provide instructions for setting up custom systems. A more recent version that includes CNN-based fly extraction will be released along with a forthcoming paper.

Some critical prerequisites:

• Python 2.7
• numpy, pandas, numba, scipy
• OpenCV > 2.4.11
• Qt and PyQt
• Panda3D

Important modules in src:

• bruchpilot/modules: Core system
  • gui.py: User interface
  • camera.py: Camera management and image processing
  • opto.py: Tools for optogenetic control
  • stimulus.py: Framework for displaying visual stimuli
  • trigger.py: Tools for triggering the camera array
  • shared.py: Inter-process communication
• bruchpilot/peripheral: Library for communicating with PointGrey cameras and the trigger system
• bruchpilot/tracking:
  • reconstruct_fast.py: Implementation of a 3D reconstruction tool based on the Hungarian algorithm (adapted from Ardekani et al., 2013), optimized via numba
  • tracker.py: Simple Kalman tracker
• scripts: Tools for calibration and post-processing of data