Filtering, Sampling, and Information Content within Satellite-Derived Multispectral or Mixed-Resolution Imagery

Abstract

The importance of an ?alias-free imaging? approach to multidimensional ?sounder? instrumentation is highlighted, together with its potential for improving noise-equivalent differential temperature and spatial resolution over current designs. Small design changes may improve postdetection signal-normalized noise spectral density by several decibels, simplify mechanical design, and facilitate composite use of instruments or channels with differing sampling and resolution. In the quest for more fundamental improvements, progress in sounder instruments has bypassed this approach; traditionally accepting scene alias, while suppressing differential-channel alias by matching and aligning channels. Newer instruments have introduced variation in alignment and channel resolution, and although relative footprint spacing has been reduced, images are insufficiently sampled to remove alias effects under difficult conditions. This study shows how alias can be eliminated without incurring unrealistic communications bandwidth and demonstrates the potentially severe penalties of failing to control alias for several model configurations similar to real instrument designs. If scientific users who specify sounder instruments adopt this approach to achieving their objectives, engineers would have the freedom to develop better and more cost-effective microwave, and especially infrared, instruments.