Occupancy analysis system using depth sensing to determine the movement of people or objects

Abstract

A depth sensing device, situated above a threshold to a room, collects asynchronous, non-personally-identifiable data regarding a count of people entering and exiting across the threshold. Phase shift of modulated infrared light reflected from a light source to an object is measured and converted into depth data indicating the depth of the pixels therein. The depth data is used to classify detected objects as human, and to track their movement bi-directionally through the threshold. A running count of humans entering or exiting the room is streamed to a remote server, which aggregates data collected from a number of depth sensing devices. The server then makes that data available at various levels of visualization corresponding to different levels of a hierarchy of nested virtual spaces, such as a multi-entrance room, or a building. Entities that access this data may see a real-time, accurate count of humans currently occupying the virtual space, as well as historical data.

Dependencies are inferred from the claim text.