Vehicle Occupant Mass Estimation via 2D/3D Pose Skeleton Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mass estimation technologies for vehicle occupants are inaccurate, bulky, costly, and not ideally suited for integration with vehicle systems, often providing wrong mass estimations due to vehicle acceleration and requiring more alignment than necessary.
Innovation Solution
A method and system using a processor to obtain 2D and 3D images of vehicle occupants, applying pose detection algorithms to generate skeleton models, filtering based on predefined criteria, and processing features to estimate mass, which can be used to control airbag deployment and other vehicle systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If prior mass estimation technologies (pressure pads or optical sensors) are used, then mass measurement function is provided, but measurement precision deteriorates due to vehicle acceleration interference
Solution Approach 1:
The patent replaces mechanical pressure sensors with an optical imaging system that captures images of the occupant and uses image processing algorithms to estimate mass. This substitution eliminates the direct mechanical contact that causes measurement errors during vehicle acceleration, as the optical system measures visual characteristics (body volume, density distribution) rather than physical pressure.
Solution Approach 2:
The system creates a visual copy (image) of the occupant's body and processes this copy to extract mass-related features. By working with image data rather than direct physical measurement, the system avoids the interference of vehicle acceleration on the measurement process while still deriving accurate mass estimates from the captured visual information.
2Adaptability or versatility
If prior mass estimation systems are implemented, then mass classification capability is achieved, but device complexity increases due to bulky components and alignment requirements
Solution Approach 1:
The patent employs a universal image processing approach that can classify occupants into mass categories (e.g., child, adult, large adult) using the same imaging and processing pipeline. The system handles multiple classification tasks and adapts to different occupant types without requiring separate specialized sensors or complex calibration procedures for each category.
Solution Approach 2:
The replacement of mechanical sensing components with an optical imaging system significantly reduces device complexity. The imaging system consists of standard cameras and processors rather than bulky mechanical sensors, eliminating alignment issues and reducing the physical space required for sensor installation while maintaining mass classification capabilities.
3Measurement precision
If prior spectrometers are used for mass estimation, then measurement function is provided, but device complexity increases due to bulkiness and alignment requirements
Solution Approach 1:
The patent replaces complex spectrometer hardware with simpler optical imaging devices (cameras). The mass estimation is achieved through image processing algorithms that analyze visual characteristics of the occupant rather than through complex spectral analysis, thereby reducing device size, eliminating alignment requirements, and lowering cost while maintaining measurement accuracy.
Solution Approach 2:
Instead of using physical spectrometers to analyze material properties, the system creates digital copies (images) of the occupant and performs virtual analysis through computer vision algorithms. This approach achieves mass estimation without the bulk and alignment sensitivity of physical spectrometric instruments.
Data Source
AI summary
There are provided methods and systems for estimating the mass of one or more occupants in a vehicle cabin comprising obtaining multiple images of the one or more occupants, comprising a sequence of 2D (two dimensional) images and 3D (three dimensional) images of the vehicle cabin captured by an image sensor, applying a pose detection algorithm on each of the obtained sequences of 2D images to yield one or more skeleton representations of said one or more occupants and combining the one or more 3D images of said sequence of 3D images with said one or more skeleton representations of said one or more occupants to yield skeleton models and analyze the skeleton models to extract one or more features of each of the one or more occupants and further process the one or more extracted features of the skeleton models to estimate the mass of the one or more occupants.