In order to validate that objects from the non-KIC catalogs appear on the Kepler detectors, we "associate" them with their nearest KIC-object neighbors. The basic idea is this: if KIC target A is calculated to be 30 pixels from the closest detector edge, and non-KIC target B is within 20 pixels of target A, then target B must be at least 10 pixels from the edge and therefore also on the detector.
Before associating KIC and non-KIC targets, the IRT entries were matched to KIS entries, IRT entries to UBV entries, and KIS entries to UBV entries. Matching was based roughly on a separation of 1 arcsec or less (1.5 arcsec for UBV). If a UBV or KIS source matched an IRT source they were considerd the same target and their data was added to the IRT entry.
Based on cross-correlating the KIC with itself, it was found that the KIC targets were almost always within 45 arcsec of each other. Therefore, using 45 arcsec (i.e., roughly 11 pixels) as a search radius, the IRT targets were found that were associated with the on-CCD KIC targets. The distances between the KIC target and the associated IRT targets were then compared to the distance between the KIC target and the closest detector edge. Subtracting the former from the latter was used as an estimate for the distance to edge value. This process was repeated for each of the 4 Kepler seasons and each of the non-matched KIS and UBV entries.
For any season in which a non-KIC target was at least 4 or more pixels from the edge, that target was considered on the detector for that season.