Real-Time Sound Source Recognition Through Acoustic Feature Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise pollution detection methods fail to accurately identify noise sources and their annoyance levels, leading to potential suspension of construction sites and financial penalties, without considering human perception.
Innovation Solution
A method and system for real-time identification of noise sources using sound sensors, preprocessing, feature extraction, and classification models like convolutional neural networks, combined with sound event detection and resident reports, to determine and notify noise nuisances.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sound level meters are installed to measure noise levels, then noise detection capability is improved, but the ability to identify specific noise sources and their annoyance levels deteriorates
Solution Approach 1:
The patent segments the noise monitoring task into multiple components: sound level measurement, sound event detection, noise source identification, and annoyance level assessment. Each component is handled by specialized algorithms processing different features of the acoustic signal, allowing simultaneous achievement of precise measurement and detailed source identification.
Solution Approach 2:
The patent transitions from traditional single-dimension noise level measurement to multi-dimensional analysis by incorporating spectral features, temporal patterns, and source classification. This dimensional expansion enables the system to identify specific noise sources and assess their annoyance levels while maintaining accurate noise level measurement.
2Device complexity
If traditional noise monitoring methods are used, then implementation simplicity is maintained, but real-time noise management and resident communication effectiveness deteriorate
Solution Approach 1:
The system performs preliminary classification of noise sources and pre-assessment of annoyance levels in real-time, enabling proactive noise management. By identifying noise sources and their potential impact before thresholds are exceeded, the system allows for preventive actions and improved communication with residents about upcoming noise events.
Solution Approach 2:
The patent implements feedback mechanisms where noise monitoring results, source identifications, and annoyance assessments are continuously communicated to stakeholders. This real-time feedback loop enables dynamic noise management decisions and improves resident communication by providing actionable information about noise sources and their impact.
Data Source
Figure 1
Figure 2
AI summary
The present invention relates to a method for identifying a sound source comprising the following steps: (S1): acquisition of a sound signal; (S2): application of a frequency filter to the acquired sound signal in order to obtain a filtered signal; (S4): extraction of a matrix of features associated with the filtered signal; (S5): identification of the source by applying a classification model to the feature matrix extracted in step (S4), the classification model having, as its output, at least one class associated with the source of the acquired sound signal.