Oliver C., Quegan S. Understanding Synthetic Aperture Radar Images

Издательство SciTech Publishing, 2004, -520 pp.The authors have been collaborating on SAR research since 1982, when Shaun Quegan joined Marconi Research Centre, Great Baddow, and Chris Oliver was at the Royal Signals and Radar Establishment, Malvern (later to become part of the Defence Research Agency (DRA), which is now the Defence Evaluation and Research Agency (DERA)). At that time the status of SAR as a means of producing images of the Earths surface was very different from the present. Only the shortlived Seasat mission had provided spaceborne data, and few airborne systems were in operation. Images were not widely available (although the DRA X-band SAR provided us with a data source) and in most cases suffered from various types of defect. Image quality issues often overshadowed the task of information extraction so that, in common with many other researchers, we invested considerable time and effort on methods of data correction. The success of this work (such as the signal-based motion compensation schemes described in Chapter 3) in combination with significant progress in system engineering and calibration has ensured that close to ideal quality SAR data is now to be expected from modern systems. We have also seen the development of new techniques, such as polarimetric and interferometric SAR, and the advent of a diversity of SAR data providers, including spaceborne systems (ERS-I and 2, JERS, and Radarsat) and flexible multifrequency polarimetric systems carried on aircraft or the Space Shuttle.
As the accent has changed from data correction, the information contained in SAR images has come to occupy its rightful place as the purpose of data acquisition. The investigation of methods for information extraction has formed the second and continuing strand in our long collaboration. In this we have been much helped by the complementarity of our interests. Chris Oliver has primarily been concerned with the development of image-interpretation tools (usually for high resolution SAR in a military context), while Shaun Quegan has concentrated more on their impact on remote sensing applications. Necessarily, there has been significant overlap in our work. This combination of interests means that we can go some way towards crossing the cultural divides that hamper SAR development, in particular those that separate military from civil applications and applications from algorithm development.
Military applications are primarily concerned with detecting and recognizing targets, which usually demands imagery of the highest resolution. Civil applications, on the other hand, may require information about many diverse aspects of the environment, normally at lower resolution. The two have common ground in the need to understand the properties of distributed scatterers: in the military case because this constitutes the background clutter against which target detection takes place; in the civil case because this clutter is often the primary object of interest. More generally, both types of application have an interest in scene understanding. However, military needs, where cost may not have been a dominating issue, have given rise to sophisticated techniques that are potentially of considerable value to the civil field. These have become accessible due to the reduction in the cost of computing. Such is true of many of the developments described in this book.Principles of SAR Image Formation
Image Defects and Their Correction
Fundamental Properties of SAR Images
Data Models
RCS Reconstruction Filters
RCS Classification and Segmentation
Texture Exploitation
Correlated Textures
Target Information
Information in Multichannel SAR Data
Analysis Techniques for Multidimensional SAR Images
Classification of SAR Imagery
Current Status and Future Prospects