The ITF originally acquired for the LWP camera in 1978-1980 was a ``mini-ITF'' with one image for most of the levels. These images were taken under widely varying spacecraft conditions, with some images taken during optimization testing when camera parameters were deliberately modified. The ITF taken in 1984 was taken under controlled conditions and the images are geometrically similar, allowing the construction of a raw space ITF. However, it was discovered during testing of this ITF that the LWP dataset of science images divide generally into two subgroups. Images taken before 1984 tend to have similar geometric distortions to one another, and images taken after 1984 tend to have similar geometric distortions, but the two subgroups are not geometrically similar to each other. Clearly, there was a discrete change in the camera operating parameters in 1984, possibly in the voltages set in the camera, but the information available via telemetry is not sufficient to identify the exact nature of the change. Unfortunately, images taken before the 1984 ITF (a very small percentage of the total LWP dataset) are geometrically similar to the LWP ITF (Epoch 1984). Images taken after the 1984 ITF are not well aligned with the 1984 ITF. This situation prompted the acquisition of a new ITF for the LWP camera in 1992.
The LWP ITF acquired in 1992 is virtually always better aligned in raw space with the science images acquired since 1984 than the 1984 ITF. However, the bimodal nature of the image distortion for the LWP camera manifested itself in the null images acquired for the 1992 ITF. The null level images formed two discrete groups. Approximately half the null images are geometrically similar to each other and form an internally consistent group (Null A). The other group of null images (Null B) are geometrically similar to each other in raw space, but the two groups are not sufficiently aligned geometrically to be co-added without loss of accuracy in PHOTOM. While somewhat more than half of the science images acquired with the LWP camera are better aligned with Null B than Null A, the intensity level of Null B is higher than that of Null A, resulting in the science data not being fully bracketed by the calibration ITF data in intensity when Null B is used. The Final Archive processing uses the 1992 LWP ITF constructed with Null A.