As is the case with the LWP camera, two ITFs were acquired for the LWR. The first ITF (1978 epoch) was acquired with no constraints placed on camera temperature (THDA). Hence, a viable raw space ITF could not be constructed from the 1978-epoch dataset (i.e., the individual UV-Flood images in a given ITF level did not align geometrically with one another). The second ITF (1983 epoch) was taken under more controlled spacecraft conditions; however, an ITF constructed purely using the 1983-epoch images (LWR83R94A or ITF A) did not correlate well with pre-1984 science images. As a result, a ``hybrid'' 1983-epoch ITF (LWR83R96A or ITF B) was generated where all levels (except the null) of the ITF were constructed from UV-Flood images taken during the 1983 ITF acquisition and the null level is composed of six null images taken during the 1978-1983 time period. LWR science images show a marked bimodality as far as ITF preference is concerned. For the most part, the majority of low-dispersion images taken prior to 1984 tend to correlate well with ITF B; while images taken after this date register well with the ITF A.
LWR ITF selection in NEWSIPS low-dispersion processing is made on an image-by-image basis. Each low-dispersion LWR image is processed twice through the image registration (CROSS-CORR) step; once using ITF A and another time utilizing ITF B. NEWSIPS then chooses the ITF which yields a higher median cross-correlation coefficient for subsequent processing of the image. High-dispersion LWR images do not go through such a selection process; only ITF A is used. This is because no clear improvement in image quality was seen when using ITF B over ITF A in high dispersion.