Conveyance Gesture Control via Depth-Color Fusion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing conveyance devices, such as elevators, face challenges in accurately capturing passenger behaviors due to limitations in current 2D and 3D sensors, which struggle to distinguish between passengers and background objects, especially in shadow regions or when passengers wear similar clothing, and fail to provide sufficient resolution for explicit and implicit gesture recognition.

Innovation Solution

A system that combines structured-light based 3D depth sensing with video streams to recognize gestures, using an array of sensors to cover a wide field of view and provide high spatial resolution, allowing for accurate detection and classification of gestures by fusing depth and color information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If 2D sensors operating on color or intensity information are used, then the system can detect passenger presence, but the sensors cannot distinguish between passengers wearing similar colored clothing or discriminate between passengers and background objects of similar color

Engineering Contradiction:
Improvepassenger detection accuracyVSAvoidpassenger classification reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent transitions from 2D color/intensity sensing to 3D depth sensing by introducing a time-of-flight camera that measures the time for light to travel to and from objects. This adds a temporal dimension to the sensing process, enabling depth measurement and creating a three-dimensional representation of the environment that resolves ambiguities in two-dimensional color-based detection

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Loss of information

If 3D sensors providing depth information are used, then the system can capture spatial information, but the sensors cannot generate depth estimates in shadow regions due to distance differences between emitter and receiver

Engineering Contradiction:
Improvedepth information coverageVSAvoiddepth measurement accuracy
Core Design Contradiction:
Loss of informationVSMeasurement precision

Solution Approach 1:

The patent combines multiple sensing modalities - time-of-flight depth sensing, stereo vision, and traditional 2D imaging - into a unified sensor system. By merging these complementary technologies, the system can cross-validate depth measurements and compensate for the weaknesses of individual sensors in shadow regions, achieving more complete and accurate three-dimensional reconstruction

Inventive Principle:
Principle #5Merging (Combining)

3Adaptability or versatility

If traditional sensors are used for conveyance control, then the system can detect basic passenger presence, but the sensors are unable to capture many passenger behaviors including explicit and implicit gestures

Engineering Contradiction:
Improvegesture recognition capabilityVSAvoidbehavior detection precision
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent implements dynamic gesture recognition by continuously capturing sequences of three-dimensional data over time and analyzing temporal patterns of motion. The system adapts to different gesture types and passenger behaviors by processing dynamic changes in depth information, enabling recognition of both explicit intentional gestures and implicit behavioral patterns

Inventive Principle:
Principle #15Dynamics

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

Enables precise and reliable gesture-based control of conveyance devices, improving passenger experience by accurately sensing gestures over a wide range and field of view, enhancing elevator operation and security monitoring.

Implementation Method 1

3D sensors that provide depth information may be unable to generate an estimate of depth in a so-called 'shadow region' due to a difference in distance between an emitter/illuminator (e.g., an infrared (IR) laser diode) and a receiver/sensor (e.g., an IR sensitive camera)

Methodology Applied
Scientific EffectTime of flight: Time of Flight

Data Source

PatentEP3033287B1Sensors for conveyance control
Publication Date: 2024.10.09 OTIS ELEVATOR CO
  • EP3033287B1 patent drawingFigure 1
  • EP3033287B1 patent drawingFigure 2~3
  • EP3033287B1 patent drawingFigure 4

AI summary

A method includes generating a depth stream from a scene associated with a conveyance device; processing, by a computing device, the depth stream to obtain depth information; recognizing a gesture based on the depth information; and controlling the conveyance device based on the gesture.