Although the scene construction process can vary depending on the type of scene to be built, the purpose that it is being built for and the scale of the scene, the process can generally by summarized by the following steps:
Identify area to model. The scene to be constructed may be a real area or it may be a fictional location. If the candidate scene is a real scene, it is helpful to acquire ground and overhead imagery of the area to help construct features.
Locate or generate terrain data for the area. For real scenes, this may be in the form of a U.S.G.S Digital Elevation Model (DEM). For a fictional scene, this could imply "designing" the general terrain and features that it will include.
Create a facet model of the terrain. This might entail using a triangular irregular network (TIN) generation program that can process a digital elevation model. For a fictional scene, it might imply hand creating a terrain that matches the desired features of the proposed scene.
Construct 3D models of the various objects in the scene. This would include houses, trees, buildings, vehicles, etc. This can be accomplished in a variety of a Computer Aided Design (CAD) applications including AutoCAD and Rhinoceros.
Acquire the necessary material thermodynamic and optical properties for the objects in the scene. This includes acquiring spectral reflectance and emissivity curves (or reusing data from other DIRSIG scenes). Spectral data from other sources will need to be converted to the appropriate DIRSIG file format (see the Spectral Emissivity File and Spectral Extinction File sections).
Create a DIRSIG Material Database File for your scene. If the scene is similar to an existing scene, the Material Database File for that scene can be used as a base. In some cases, a duplicate copy material database from another scene can be used.
Create a material map for the scene. This is an optional but powerful technique that can be used to attribute the facetized terrain. If aerial imagery is available, it might be used as an input to terrain classification algorithm. See the Property Mapping Techniques chapter to learn more about this process.
Import and attribute the CAD objects using the bulldozer tool in "Object Mode" (see The Bulldozer User Manual). This tool will currently import geometry using the Alias/Wavefront OBJ 3D graphics format. When the model is attributed, it is saved as a Geometric Database File.
Create any texture maps that can be used to drive in-material spatial-spectral variation. Using the DIRSIG texture algorithm requires that you have overhead imagery in at least one spectral channel and that the Spectral Emissivity File for the corresponding material contains many curves. This process is described in better detail in the Property Mapping Techniques chapter.
Use the bulldozer tool in "Scene Mode" to place the attributed objects (GDB files) onto facetized terrain. If aerial imagery is available, these objects can be placed with the aid of "tracing paper" (refer to the The Bulldozer User Manual).
Create a DIRSIG configuration file (see the Input Configuration File section) to simulate a sensor imaging your scene.
Run the DIRSIG simulation and review the generated imagery.