UC Santa Cruz

Due to the lack of visual cues and access to detailed environmental information, many visually impaired people are reluctant to travel independently in unfamiliar locations. Several lines of research have addressed the problem of wayfinding for blind individuals, mainly focusing on localization and guidance systems. Even when these systems are available, pre-journey learning is a valuable resource, helping blind people to create a mental image of their surrounding and to maintain their orientation in the case of system malfunction.

In this dissertation, I introduce an end-to-end systematic tool for a collective spatial mapping of an indoor environment consisting of small-scale features (i.e., building fixtures, room furnitures, types of floor-covering, etc), and for the production of tactile maps at multiple scales, promoting pre-journey spatial awareness. This systematic application, named Semantic Interior Mapology or SIM, is a three-fold. The first component, Map Conversion Tool, allows one to quickly and accurately trace a floor plan from an architectural image of it. Next, the Map Population Tool component segments out small-scale elements of interest from a 3D scan of a room, and then geo-registers them within the building's spatial layout. These initial semantic spatial relationship is stored as a vectorized format that is amenable to reproduction in multiple modalities. The last component, Map Authoring Tool, produces on demand the tactile maps of indoor environments from the building's structural layouts and its 3D-scanned interior spatial contents that are captured and vectorized previously. Such maps are embossed at different spatial scales, representing the building's general structure, a zoomed-in of a specific area, or an interior of a room, with specific constraints on the density and distances of tactile features. My end-to-end systematic application described in this dissertation minimizes the time and effort required to acquire a detailed description of an indoor space, and produces accurate results even in the case of complex building layouts.