Acoustic Material Estimation Using Reverberation and Reflected Sound
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for material identification, such as those using ultrasonic waves, lack the capability to accurately estimate the material of an estimation target based on sound reflections.
Innovation Solution
A material estimation method that involves emitting an emission sound, obtaining a reflected sound, extracting acoustic features like reverberation time and frequency characteristics, and estimating the material based on these features using a learning model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional ultrasonic wave methods are used for substance identification, then the device can identify hardness based on frequency characteristics of reflected waves, but it lacks the capability to accurately estimate material type
Solution Approach 1:
The patent segments the acoustic feature extraction process into multiple distinct components: reverberation time extraction, frequency characteristics extraction, and spectral centroid calculation. Each segment processes specific aspects of the reflected sound wave independently, allowing comprehensive material characterization without information loss.
Solution Approach 2:
The patent transitions from conventional one-dimensional frequency analysis to multi-dimensional acoustic feature analysis by incorporating time-domain (reverberation time), frequency-domain (frequency characteristics), and spectral-domain (spectral centroid) features simultaneously, enabling accurate material type estimation.
2Measurement precision
If only frequency characteristics of reflected ultrasonic waves are used, then the identification process is simple, but the material estimation accuracy is insufficient
Solution Approach 1:
The patent merges multiple acoustic feature extraction methods (reverberation time analysis, frequency characteristics analysis, and spectral centroid calculation) into a unified material estimation system. The learning model integrates these combined features to achieve accurate material identification while managing system complexity through systematic organization.
Solution Approach 2:
The patent introduces a learning model as an intermediary component that processes the complex multi-dimensional acoustic features and transforms them into accurate material estimations. This intermediary handles the complexity of feature integration while providing simplified material classification output.
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 materials by comparing acoustic features with stored data, enhancing the accuracy of material identification.
Implementation Method 1
obtaining a reflected sound of the emission sound reflected by the estimation target
Implementation Method 2
extracting an acoustic feature of the estimation target based on the reflected sound, the acoustic feature including information pertaining to a reverberation time of an impulse response
Data Source
AI summary
A material estimation method includes emitting an emission sound toward an estimation target of which a material is to be estimated; obtaining a reflected sound of the emission sound reflected by the estimation target; extracting an acoustic feature of the estimation target based on the reflected sound, the acoustic feature including information pertaining to a reverberation time of an impulse response; and estimating the material of the estimation target based on the acoustic feature of the estimation target.