Air conditioner production line abnormal sound detection method and system
By combining a microphone spherical array and a panoramic camera with the HOA-SHT domain analysis framework and convolutional neural network, we have achieved efficient and accurate detection of abnormal noises from air conditioner wall units, solved the problem of foreign object noise in the impeller blades, and improved production efficiency and user experience.
Patent Information
- Application Number
- CN202510466688.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-15
- Publication Date
- 2025-10-24
- Estimated Expiration
- 2045-04-15
AI Technical Summary
In the current production of wall-mounted air conditioners, noise problems caused by foreign objects on the surface of the fan blades are difficult to detect efficiently and accurately through manual listening, resulting in high false detection rates, high missed detection rates, and low efficiency.
By combining a microphone spherical array and a panoramic camera with the HOA-SHT domain analysis framework and convolutional neural network, a panoramic sound image is generated. Sound signals are extracted through multimodal localization and multi-scale Mel spectrum, and abnormal noise is detected using a convolutional neural network, thus achieving efficient and accurate classification of abnormal noises from air conditioners.
It significantly improves the accuracy and efficiency of detecting abnormal noises in wall-mounted air conditioners, reduces false detection and missed detection rates, and improves production quality and user experience.