Efficient and automatic generation of three-dimensional models
Lead Inventors: Peter Allen and Michael ReedProblem or Unmet Need:Automatic three-dimensional construction of a computer representation of an object or a scene is becoming increasingly important for applications using computer modeled objects. Conventional techniques which construct three-dimensional images suffer from several limitations; these include (1) the need for many scans from many angles around the object to achieve a representation of sufficient resolution, (2) the need to move the relative position of scanner so that every surface of the object is scanned with overlapping views, and (3) the inability to adjust angles to image occlusions accurately. As a result, conventional techniques for three-dimensional modeling may produce images of insufficient resolution and require long processing times which prevent (among other uses) the ability to perform real time imaging.This technology consists of a system and method for efficiently generating a three-dimensional representation of an object or a scene. The technology minimizes the number of required scans needed to create a high-quality representation at a desired resolution. During the creation of the initial representation (a solid model), occlusion tags mark surfaces which need further imaging from a different perspective. A planning stage determines which view will include the most surfaces tagged "occluded" by using a volume visibility technique. The results of the subsequent scan are combined with the current representation to form a more accurate composite representation, and imaged surfaces (or surfaces which are now not occluded) do not require further scanning, thereby minimizing the number of required scans and the processing time.
Fast and simple with correct model at each step Can tune model fidelity to specifications of application Reduces minimum number of scans needed
Rapid prototyping Advanced structural analysis Manufacturing design and process planning (industrial applications) Population of realistic, high fidelity, virtual worlds (e.g. Google maps, movies) Medical imaging
This technology consists of a system and method for efficiently generating a three-dimensional representation of an object or a scene. The technology minimizes the number of required scans needed to create a high-quality representation at a de...
美國
