Backward propagation algorithms for image reconstruction: Signal processing, algorithm architecture, and applications

Abstract

Traditionally, image reconstruction algorithms were developed, following the design and configurations of the data-acquisition systems. So, the algorithms are typically special-purpose and system-specific. Consequently, there is a lack of consistency in terms of algorithm architecture, organization, and performance. In this paper, we present a unified framework for algorithm design and development. This allows us to implement image reconstruction for various data-acquisition configurations including active or passive sensing, linear or circular receiving apertures, CW or wideband illumination, and monostatic or bistatic formats, based on a single theoretical framework in an organized manner. The computation schemes for both linear and circular apertures will be discussed in detail. The layered backward propagation technique, as the main processing modality, provides the flexibility for dynamic updating for changes of propagation parameters. In addition, we illustrate parallel processing and recognition as integrated components for the algorithm structure. The presentation includes the theoretical background on signal processing for image formation, and overview of algorithm architecture for various configurations, a discussion on computation complexity and commonality, and several applications.