KEYNOTE – Robert Erdmann
summary
Captures Rembrandt’s Night Watch at 5 micrometer resolution producing 717 gigapixels from 8,439 100MP photos, ~23 seconds per shot. Builds an orbital frame with a Hasselblad H6D plus 15 laser range sensors and a sled for XYZ+pitch+yaw control. Uses Python stack: Jupyter/ipywidgets + asyncio/websockets, scikit‑learn for robust plane fits, PyTorch for focus nets, pyvips for out‑of‑core images, JAX for fast GPU interpolation, scipy/shapely/networkx/pyamg for registration and fusion, and Argyll/color‑science to achieve Delta E < 2.3. The pipeline works end‑to‑end.
AI-written summary · about
chapters (15)
- 0:00Start
- 4:30Cad Rendering
- 5:17Rotational Degrees of Freedom
- 5:47Digital Camera
- 6:53Architecture of the System
- 7:07Control Subsystem
- 15:52Depth of Field
- 16:16Shape from Focus
- 17:21Monitor the Sharpness of the Painting
- 20:22Color Calibration
- 24:03Fiducial Marks
- 25:2720 Micron Resolution Photos
- 29:28Homography
- 31:05Jaxx
- 37:48Final Result
related talks
45mFrom Radar Echoes To Real-Time Recognition: Deep Learning on Another Planet an Open Source Workflow
31mFrom Radar Echoes To Real-Time Recognition: Deep Learning on Another Planet -An Open Source Workflow
37mUnstructured Data Processing with a Raspberry Pi AI Kit and Python
30mAI On-board of Satellites, Towards Autonomous Scientific Instruments
38mEldercam: Event-driven AI for Surveillance
45mStars, Planets, and Python
41mKEYNOTE: Image processing, AI, and autonomous systems
29mReconstructing Unseen Dimensions in Biophotonics Through Python