Acoustic Work Estimation System Using Inaudible Sound Reflection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional surveillance systems that use cameras to monitor work sites often infringe on privacy and suffer from decreased accuracy due to ambient brightness changes, making it difficult to estimate the content of work performed by individuals while protecting their privacy.
Innovation Solution
A work estimation method and system that utilizes sound information, specifically reflected inaudible frequency range sounds and work sounds, to estimate the content of work performed by a person, using trained models to output image, tool, and work information, thereby protecting privacy and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If camera-based surveillance systems are used to monitor work sites, then work content can be visually captured, but privacy infringement occurs and accuracy decreases due to ambient brightness changes
Solution Approach 1:
The patent replaces optical surveillance systems (cameras) with acoustic sensing systems (microphones) to detect work content. Sound waves in the inaudible frequency range are used to capture work site activities without the privacy concerns and environmental sensitivity of visual systems, thereby substituting a mechanical/optical detection method with an acoustic one.
Solution Approach 2:
The patent changes the detection parameter from visible light (camera-based) to inaudible sound frequencies (acoustic waves). By utilizing sound waves outside the human audible range, the system achieves work monitoring capability while avoiding privacy infringement associated with visual surveillance and eliminating sensitivity to ambient brightness conditions.
2Measurement precision
If camera-based surveillance systems are used to monitor work sites, then work content can be visually captured, but accuracy decreases due to ambient brightness changes
Solution Approach 1:
The patent replaces optical surveillance systems (cameras) with acoustic sensing systems (microphones) to detect work content. Sound waves in the inaudible frequency range are used to capture work site activities without the privacy concerns and environmental sensitivity of visual systems, thereby substituting a mechanical/optical detection method with an acoustic one.
3Object-affected harmful factors
If sound information processing is implemented to protect privacy, then privacy is maintained, but system complexity increases due to multiple trained models
Solution Approach 1:
The patent segments the work estimation task into three distinct neural network models, each specialized for a specific function: one model estimates work area from reflected sound, another estimates tool type from work sound, and a third estimates work content by combining both inputs. This segmentation allows each model to be optimized for its specific task while maintaining overall system privacy protection through acoustic-only sensing.
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 accurate estimation of work content while maintaining privacy, reducing noise and incorrect estimations by using sound-based data processing, and optimizing data input to improve estimation accuracy and reduce power consumption.
Implementation Method 1
first sound information related to a reflected sound that is sound obtained by reflection of emission sound in an inaudible frequency range
Implementation Method 2
second sound information related to a work sound generated by the work performed by the person
Data Source
AI summary
A work estimation method includes: obtaining first sound information related to a reflected sound that is a sound obtained by reflection of an emission sound in an inaudible frequency range and second sound information related to a work sound generated by a work performed by the person; outputting image information that indicates a work area of the person by inputting the first sound information to a first trained model; outputting tool information that indicates a tool that is being used by the person, by inputting the second sound information to a second trained model; and outputting work information that indicates the content of the work, by inputting the image information and the tool information to a third trained model.