Accomplishment:

Rewrote heatmap.py to match the final design report: 61 angle bins over ±52° (±50° drawn), the 105° camera with the angle-to-pixel map, and 4-sweep smoothing. Render time for a 1280×720 frame dropped from 45 ms to about 5 ms.
Rebuilt capture_with_heatmap.py: MJPG 1280×720 at 30 fps, only new sweeps go into the smoother, per-stage timing and a CSV log. It held 30 fps (about 33 ms per frame) on synthetic input. It also ran with display_output.py, tested against a stand-in for the PYNQ DisplayPort driver.
Added source.py (mock source, plus a real register reader that loads the overlay, checks an ID register and rejects torn reads) and calibrate_threshold.py (sets the false-alarm threshold from no-source sweeps; 0.50% observed on 17,897 mock updates). Tested the register reader against a fake PYNQ register block: 0 torn reads in about 100,000 reads.
Last week 10/10/2026: finialized all the design and drafted the design report

Schedule Status:

On track. Board integration depends on Roshni’s bitstream and register map.

Agree the register map with Roshni, get the .bit/.hwh, and start the mock-to-real data swap.
Test the Mini DisplayPort to HDMI cable against the actual LCD, and confirm the mode it supports.
Run the camera on the board at 30 fps, calibrate its field of view and offset, and calibrate the false-alarm threshold on real sweeps.