Inference-based Geometric Modeling for the Generation of Complex Cluttered Virtual Environments

摘要

As the use of simulation increases across many diff erent application domains,the need for high- fidelity three-dimensional virtual representations of real-world environmentshas never been greater. This need has driven the research and developmentof both faster and easier methodologies for creating such representations. In this research,we present two diff erent inference-based geometric modeling techniques thatsupport the automatic construction of complex cluttered environments.The fi rst method we present is a surface reconstruction-based approach thatis capable of reconstructing solid models from a point cloud capture of a clutteredenvironment. Our algorithm is capable of identifying objects of interest amongst acluttered scene, and reconstructing complete representations of these objects even inthe presence of occluded surfaces. This approach incorporates a predictive modelingframework that uses a set of user provided models for prior knowledge, and appliesthis knowledge to the iterative identifi cation and construction process. Our approachuses a local to global construction process guided by rules for fi tting high qualitysurface patches obtained from these prior models. We demonstrate the application ofthis algorithm on several synthetic and real-world datasets containing heavy clutter and occlusion.The second method we present is a generative modeling-based approach that canconstruct a wide variety of diverse models based on user provided templates. Thistechnique leverages an inference-based construction algorithm for developing solidmodels from these template objects. This algorithm samples and extracts surfacepatches from the input models, and develops a Petri net structure that is used by ouralgorithm for properly fitting these patches in a consistent fashion. Our approach usesthis generated structure, along with a defi ned parameterization (either user-defi nedthrough a simple sketch-based interface or algorithmically de fined through variousmethods), to automatically construct objects of varying sizes and con figurations.These variations can include arbitrary articulation, and repetition and interchangingof parts sampled from the input models.Finally, we affim our motivation by showing an application of these two approaches.We demonstrate how the constructed environments can be easily usedwithin a physically-based simulation, capable of supporting many diff erent applicationdomains.