Computer image generation (CIG) successfully provides images for a large proportion of visual simulation tasks, using polygon models of the environment. However, for close approach to terrain, e.g., for air/ground weapon delivery, confined area maneuvering and harbor navigation, the image detail is insufficient and the cost of generating the high detail data bases is becoming increasingly prohibitive.
The authors are developing a new form of CIG - the Real Environment Algorithm for Line Scanning (Realscan) System, using a uniform grid type digital data base of contoured, textured terrain semi-automatically generated from Defense Mapping Agency (DMA) and stereo photo data. The system will use video disc bulk storage, potentially visible data (for a given position and attitude of the simulated vehicle) being read into random access memory and then addressed in accordance with an algorithm giving correct perspective mapping into the display in real time with automatic elimination of hidden areas of terrain and without aliasing (sampling artifacts).
The image detail in a system of this type can be limited only by the display resolution, giving a very large increase in scene detail over what is likely to be available for many years with polygon modeled CIG. The application of the system to IR is being studied with the aim of producing coordinated visual/IR/radar displays generated from a common data base.