In this chapter we introduce the concept of a common representational format. This is a basic concept in all multi-sensor data fusion systems and is concerned with the conversion of all input signals to a common format. The process of converting sensor observations into a common representational format involves spatial, temporal and semantic alignments together with radiometric normalization. The chapter contains introductory discussion of these functions and includes more than fifteen extended examples of different common representational formats and their use in practical data fusion applications. The examples include the use of “off-the-shelf” deep-learning features.