2001AJ....121.2239G

Speckle interferometric observations of double stars enjoy considerably higher weight in orbital solutions than visual observations. This is justified on the basis of their higher precision and impersonal nature. It is crucial, therefore, that sources of systematic error in speckle observations are identified and controlled to a level commensurate with their low random errors. Accurate calibration is of primary importance. Other sources of systematic error that affect ICCD-based speckle interferometry include imperfect CCD charge transfer efficiency, fixed pattern noise, and nonuniform background in the autocorrelation function. All of these errors and their remedies are discussed. The restoration of observations compromised by fixed pattern noise is evaluated, and results for 30 such observations are given.