Generalized Scene Reconstruction
This work addresses the challenge of reconstructing complex real-world scenes for applications in mobile devices, AR, and drones, representing an incremental advancement in scene reconstruction techniques.
The paper tackles the problem of reconstructing 'generalized scenes'—boundless spaces with complex material properties like non-Lambertian and textureless matter—by introducing a new passive approach called Generalized Scene Reconstruction (GSR), which uses a plenoptic octree data structure to enable efficient light and matter field reconstruction on devices like mobile phones and AR glasses, and demonstrates it by reconstructing highly reflective, hail-damaged automobile body panels with a prototype imaging polarimeter.
A new passive approach called Generalized Scene Reconstruction (GSR) enables "generalized scenes" to be effectively reconstructed. Generalized scenes are defined to be "boundless" spaces that include non-Lambertian, partially transmissive, textureless and finely-structured matter. A new data structure called a plenoptic octree is introduced to enable efficient (database-like) light and matter field reconstruction in devices such as mobile phones, augmented reality (AR) glasses and drones. To satisfy threshold requirements for GSR accuracy, scenes are represented as systems of partially polarized light, radiometrically interacting with matter. To demonstrate GSR, a prototype imaging polarimeter is used to reconstruct (in generalized light fields) highly reflective, hail-damaged automobile body panels. Follow-on GSR experiments are described.