Depth Sensing Occupancy Analysis Using Phase Shift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing people counting solutions struggle to balance accuracy and privacy, often relying on personally-identifiable information or sacrificing accuracy for anonymity.
Innovation Solution
A depth sensing device positioned above a doorway uses phase shift data from modulated light to generate depth data, classifying objects as human subjects while maintaining anonymity, and processes this data locally without storing or transmitting personally-identifiable information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical cameras or badge/fob data are used for people counting, then measurement precision is improved, but personally-identifiable information is collected sacrificing privacy
Solution Approach 1:
The patent extracts only the essential counting information from the data stream while discarding personally-identifiable information. The system processes camera images to detect and count individuals but deliberately excludes facial recognition or other identifying features, thereby achieving accurate people counting without privacy intrusion.
Solution Approach 2:
The patent introduces an intermediary processing layer between the camera and the counting output. This intermediary system uses anonymous identifiers and aggregated data processing to bridge the gap between visual detection and privacy-preserving counting, allowing accurate measurement without exposing individual identities.
2Object-affected harmful factors
If thermal cameras or motion sensors are used for anonymous counting, then privacy is protected, but measurement precision and detection accuracy deteriorate
Solution Approach 1:
The patent merges multiple sensing modalities including optical cameras, thermal sensors, and motion detectors into a unified system. By combining these different sensor types, the system achieves both the privacy protection of anonymous sensing and the detection accuracy of multi-modal data fusion, overcoming the limitations of using any single sensor type alone.
3Object-affected harmful factors
If manual observational studies are used for people counting, then privacy is maintained, but productivity and scalability are reduced
Solution Approach 1:
The patent implements a self-service automated counting system that operates without human intervention. The system automatically captures, processes, and counts individuals using camera and sensor data, eliminating the need for manual observers while maintaining privacy through anonymous processing methods, thereby dramatically improving productivity and scalability.
4Measurement precision
If depth sensing with phase shift analysis is used, then measurement precision and anonymity are both improved, but device complexity increases
Solution Approach 1:
The patent employs periodic modulation of light sources and corresponding phase shift analysis to enable depth sensing. By using periodic signals rather than continuous illumination, the system achieves accurate depth measurement and object classification while reducing overall system complexity through efficient signal processing techniques.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The solution provides accurate, real-time people counting that is both anonymous and scalable, addressing the conflict between data accuracy and privacy concerns.
Implementation Method 1
A depth sensor identifies a phase shift of modulated light reflected from a light source to an object
Data Source
AI summary
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.