Neo: Virtual Object Modeling using Commodity Hardware

Abstract

Recent developments in augmented reality technology have paved way for newapplications in a wide range of areas. These include the commercial markets,medicine applications, military applications and education. The technology pro-vides immersive images to enhance our perception of the world. Augmentedreality addresses challenges related to problem-solving by seamlessly integrat-ing digital images into real-world images.In the context of construction and maintenance industry, project inspections canbe time-consuming and tedious. These inspections involve usages of expensiveand specialized hardware. Some inspections even use physical blueprints anddrawings along with standardized measurement tools. This approach can posepractical challenges and be prone to errors.In this thesis we present Neo, a surface reconstruction system on commodityhardware. It utilizes augmented reality technology by scanning physical sur-roundings and reconstructs them as virtual objects. They are displayed on topof the camera’s live preview of the real world. By using a pipeline architecturewe model the physical surroundings in terms of their shapes and visual appear-ances. Cyber-physical information about the reconstructed virtual models areannotated in real-time. Evaluations of the system show us potentials to createrealistic copies of physical objects