3D Human Tracking with Shelf Cameras and Ultrasonic Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing human tracking systems in retail environments face inaccuracies in crowded scenarios and fail to accurately recognize purchase behaviors due to reliance on two-dimensional image information and inaccurate position detection by light curtains or infrared sensors.
Innovation Solution
A method and apparatus that utilize multiple photographing devices, including at least one above a shelf, to capture comprehensive image information and combine it with non-visual sensor data, such as light curtain and infrared sensor detection, to improve human tracking accuracy and motion recognition precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multi-channel cameras are uniformly arranged for human tracking, then the system can capture two-dimensional image information, but the tracking accuracy deteriorates in crowded retail scenarios
Solution Approach 1:
The patent transitions from two-dimensional image-based tracking to three-dimensional spatial tracking by introducing depth information through time-of-flight cameras and ultrasonic sensors. This dimensional enhancement allows accurate tracking in crowded scenarios by providing spatial depth data that distinguishes overlapping individuals in 3D space rather than 2D projections.
Solution Approach 2:
The patent combines multiple sensing modalities including time-of-flight cameras, ultrasonic sensors, and traditional image cameras into a unified tracking system. This multi-sensor fusion approach merges visual, depth, and acoustic information to achieve robust tracking accuracy that compensates for the limitations of any single sensor type in crowded environments.
2Extent of automation
If light curtains or infrared sensors are used to acquire hand position, then purchase behavior detection is enabled, but the position accuracy deteriorates causing imprecise motion recognition
Solution Approach 1:
The patent merges time-of-flight camera depth data with ultrasonic sensor position detection to create a hybrid hand tracking system. The time-of-flight camera provides accurate 3D hand position and gesture information, while ultrasonic sensors detect hand movements through air coupling, compensating for each other's limitations and achieving precise hand position accuracy for purchase behavior detection.
Solution Approach 2:
The patent introduces ultrasonic waves as an intermediary medium to detect hand positions and gestures. Ultrasonic sensors emit sound waves that reflect off the hand, providing contactless position detection that complements optical methods and improves accuracy in scenarios where light-based sensors alone fail to capture precise hand movements.
3Speed
If real-time recognition is performed continuously, then immediate action detection is achieved, but resource utilization deteriorates due to unnecessary processing during inactivity
Solution Approach 1:
The patent implements periodic sampling of sensor data at optimized intervals rather than continuous processing. The system periodically checks for changes in hand position, gesture patterns, or purchase intent signals, and only triggers full recognition processing when changes are detected, reducing computational resource utilization while maintaining appropriate response times for action detection.
Solution Approach 2:
The patent performs preliminary filtering and preprocessing of sensor data to detect potential actions before full recognition processing. The system monitors for preliminary indicators such as hand approaching the shelf, changes in grip patterns, or deviations from normal browsing behavior, and only initiates comprehensive action recognition when these preliminary conditions are met, optimizing resource utilization.
Data Source
Figure 1
Figure 2
Figure 3~5
AI summary
Embodiments of the present invention provide a method for human motion analysis, an apparatus for human motion analysis, a device, and a storage medium. The method includes: acquiring image information captured by a plurality of photographing devices, where at least one of the plurality of photographing devices is disposed above a shelf; performing human tracking according to the image information captured by the plurality of photographing devices, and determining position information in space of at least one human body and identification information of the at least one human body; acquiring, according to the position information in space of a target human body of the at least one human body, a target image captured by the photographing device above a shelf corresponding to the position information; and recognizing an action of the target human body according to the target image and detection data of a non-visual sensor corresponding to the position information. Embodiments of the present invention improve the accuracy of human tracking and the precision of the human action recognition.