Image Simulation for the NEO Surveyor
Jason Surace1, Xiang (Cate) Liu1 
1IPAC/Caltech, Pasadena, CA

We describe the science image simulator for the Near-Earth Object Surveyor Mission. The purpose of the simulator is to produce realistic raw image data, with properties similar enough to real instrument data as to exercise all components of the survey data processing system, and sufficient to characterize the likely performance of the processing algorithms relative to the mission requirements. The simulator is intended to support testing of data processing algorithms and system throughput.  It is not intended for forward modeling and comparison to real data, or any similar procedure.

The simulator begins with the planned spacecraft pointing centers, expanding these to every planned detector image. The simulator then creates for each image a noiseless calibrated representation of the sky called the “astrophysical scene”. A 2-dimensional map of the zodiacal background, cosmic infrared background, and interstellar medium are added. Actual astrophysical sources derived from the ALLWISE Source Catalog (corrected for GAIA proper motion), in addition to much fainter artificial sources derived from deep Spitzer surveys, are injected as point sources. These point sources are modeled using a spatially variant PSF derived from instrument optical modeling, and include the effects of focal plane distortion. Extended sources such as large galaxies are added as either simple gaussian models or as 2-D images derived from WISE. Moving objects such as planets, comets, and asteroids (the latter derived from the NEOS Reference Small Body Population Model) are injected at the correct positions for the time of the simulated observation.

The finished astrophysical scene is then passed to a simulation of the detector readout. The model includes the gain, read noise, flatfield, dark current, cosmic rays, non-linearity, saturation and image persistence. The actual readout pattern of destructive and non-destructive reads is simulated. Upon completion, the finished image is packaged along with simulated header data to mimic the spacecraft raw data product.

The simulator is massively parallel, allowing simulated datasets as large as a full month of spacecraft operations to be computed quickly. The simulator is deployed both at IPAC and on the Amazon AWS cloud. It's modular design makes updating the detector, instrument and survey strategy easy as new test data become available. This feature makes the simulator a key part of the NEO Surveyor systems engineering toolkit to validate overall system performance, and to support trade studies when used in concert with the NEO Surveyor Survey Simulator and Survey Data System.