Abstract spatial hybrid systems for image processing

Model-based design and development has become a popular paradigm for developing embedded systems. Many modern embedded systems have components that are dedicated to capturing and processing images for object recognition. Examples of such systems include autonomous robots, spacecrafts and security devices. Such image processing functionalities are performed by software that involve nonlinear operations. Hybrid systems are models of nonlinear systems that involve the interaction of continuous and discrete dynamics. They have been used to solve problems in diverse domains ranging from operations research, design of embedded systems to the design and analysis of biological systems. In particular, they have been used as models for model-based design, development and analysis of embedded and biological systems. We introduce a new version of hybrid systems called the abstract spatial hybrid systems (machines). We describe how such systems can be used as models for image processing tasks like image registration.