|
Ruda designs high performance imaging systems to meet difficult mission requirements, but these nearly diffraction limited systems often have small margins between the image quality of the nominal design and the required performance of the as-built system. Due to this we may spend significant resources designing and operating specialized test setups to ensure that results of MTF and Ensquared Energy (EE) measurements are well-calibrated and accurate. Alternatively, wavefront measurements – like those captured by wavefront sensors and interferometers – can be taken of the system to characterize the quality of the as-built system. Wavefront measurements are typically higher resolution, faster to setup, and quicker to measure than image quality metrics, making them particularly attractive for use when validating as-built system quality. Since the wavefront is related to the point spread function, and thereby the image quality, different wavefront measurements can contain information about the system MTF and EE. Thus, if the relationship between the wavefront and image quality metrics of interest can be established for an as-built system, it is possible to supplement or fully validate MTF and EE requirements from wavefront measurements. To investigate this relationship, we used Zemax OpticStudio to generate toleranced Monte Carlo trials of two nearly diffraction limited imaging systems designed by Ruda. The Monte Carlo models were then analyzed to form large data sets for statistical analysis. For wavefront data, the simulation produces single pass and double pass wavefront Zernike decompositions as well as wavefront root mean squared error over a range of object fields and visible wavelengths. For image quality data, the MTF at three spatial frequencies and the EE at two integration lengths are computed for the same fields and wavelengths as the wavefront data. These data sets are then processed to demonstrate that high degrees of correlation can exist between wavefront data and image quality metrics in toleranced high performance imaging systems, even when there is a difference in wavelength between the metrics. Sources of noise in these correlations are identified, and paths for supplementing or validating image quality requirement with correlated wavefront measurement data through machine learning are discussed.
|
