Simulated images for an HWO survey: a yield-to-image dataset (Spohn & Stark, SPIE 2026)

Two independently scheduled multi-year simulated surveys of the same star sample for the Habitable Worlds Observatory (HWO), rendered to the image level with coronagraphoto, released to accompany Spohn & Stark, SPIE 2026. One survey is scheduled by EXOSIMS (a Python-based yield calculator) and one by the Altruistic Yield Optimizer (also called AYO). Both share the same coronagraph design and instrument parameters, so they differ primarily in the scheduler that orders the observations. A single configuration drives the yield calculators and the image-simulation pipeline together, so the same design and target list set both the observing schedule and the per-visit images. The release totals 1972 survey frames across the two surveys, together with a high-fidelity solar-system-twin stack rendered at full time and wavelength resolution (six per-channel, per-roll FITS cubes at five-minute cadence in three spectral channels). Each survey frame is a multi-extension FITS file holding the two telescope-roll images (plus and minus 22.5 degrees) for angular differential imaging, 256 by 256 pixels at 3.94 mas per pixel, with every observing parameter recorded in the FITS header. These frames are idealized pipeline demonstrations of an early HWO design (EAC-1, a charge-6 amplitude-apodized vortex coronagraph on a 7.2 m primary): they include photon noise but no speckles and no detector noise, so the signal-to-noise values are upper bounds rather than predictions of HWO performance. The per-planet ground-truth catalog is deliberately withheld so the survey can serve as a blind post-processing challenge. To compare recovered catalogs against the ground truth, contact Corey Spohn ([e-mail]). See the included README for the full contents, header-card list, parameters, and caveats.