Chapter 16 describes a new technique for extracting patterns and relations visually from multidimensional binary data. The proposed method relies on monotone structural relations between Boolean vectors in the n-dimensional binary cube and visualizes them in 2-D as chains of Boolean vectors. Actual Boolean vectors are laid out on this chain structure. Currently the system supports two visual forms: the multiple disk form and the “Yin/Yang” form.
Part 5 concludes the book with geospatial data analysis, decision making and problem solving and consists of chapter 17-21. This focus is not accidental – geospatial problems are naturally visual and spatial. Chapter 17 features a general framework for combining geospatial datasets. The framework is task-driven and includes the development of task-specific measures, the use of a task-driven conflation agent, and the identification of task-related default parameters. The chapter also describes measures of decision correctness and the visualization of decisions and conflict resolution by using analytical and visual conflation agents. Finally, the chapter elaborates mathematical (geometric and topological) techniques for decision making and problem solving for combining geospatial data.
Chapter 18 addresses imagery conflation and registration problems by providing an Analytical and Visual Decision Framework (AVDF). This framework recognizes that pure analytical methods are not sufficient for solving spatial analysis problems such as integrating images. Without AVDF, the mapping between two input data sources is more opportunistic then definitive. A partial differential equation approach is used to illustrate the modeling of disparities between data sources for a given mapping function. A specific case study of AVDF for pixel-level conflation is presented based on Shannon’s concept of mutual entropy. The chapter also demonstrates a method of computation reduction for defining overlapping image areas.
Chapter 19 looks at spatial decision making and analysis, which heavily depend on the quality of image registration and conflation. An approach based on algebraic invariants for the conflation/registration of images that does not depend on identifying common points is developed. This new approach grew from a careful review of other conflation processes based on computational topology and geometry. This chapter describes the theory of algebraic invariants and describes a conflation/registration method and measures of correctness for feature matching and conflation.
Chapter 20 presents technology for conflation algorithm development with a wide applicability domain. The sequence of steps starts from vague but relevant expert concepts and ends with an implemented conflation algorithm. The generic steps are illustrated with examples of specific steps from the development history of an area-based “shape size ratio” conflation algorithm. The fundamental “shape size ratio” measure underlying the algorithm has rather strong invariance properties, including invariance to disproportional scaling.
Chapter 21 presents an Artificial Intelligence technique for generating (on-the-fly) rules of visual decision making for use by experienced imagery analysts. The chapter addresses the construction of a methodology and tools that can assist in building a knowledge base of imagery analysis. Further, the chapter provides a framework for an imagery virtual expert system that supports imagery registration and conflation tasks. The approach involves three strategies: recording expertise on-the-fly, extracting information from the expert in an optimized way using the theory of monotone Boolean functions, and using iconized ontologies to build a conflation method.