Establishment of evaluation method for the acoustic camera

TCAUV_01_TCI2024

The acoustic camera is a multimodal sensing device that visualizes the acoustic field or location of the sound sources by using an integrated sensing system combined with an image sensor and a microphone array. Such devices can be used in situations where sound provides important information, e.g., in rescue operations, monitoring the traffic noise sources in urban areas, structural health monitoring, and enhancing security systems that rely on acoustic signals. With the increasing demand for such a device in the engineering fields, the supply of acoustic camera devices is gradually increasing as well, and various sound source localization algorithms are being developed to implement this.

However, there is currently no established method to quantify and evaluate measured results because the sensor configuration techniques and localization algorithms are specialized for each device or company. Furthermore, for devices equipped with MEMS microphones, a standardized calibration procedure has not been established, leading to a lack of traceability. Therefore, the results that have been measured using an acoustic camera are only valuable as a reference, but the field of use of the acoustic camera has been limited because the absolute value of the measured result is not quantitatively reliable.

Calibration of a MEMS microphone flush-mounted on a DUT

In this project, fundamental research was conducted to establish traceability for acoustic cameras. The main purpose of the project is to develop a systematic method to evaluate an acoustic camera and a calibration method to establish traceability. First, various calibration methods for individual MEMS microphones were compared in a pressure field, free field, and diffuse field, respectively, and one can found that the effective frequency range for each sound field calibration was identified. And then, the calibration was conducted on a DUT containing a digital MEMS microphone array, so that factors for calibration and correction for individual items could be extracted. Also, a test procedure was developed to evaluate the consistency of camera images and localization results.

System configuration for a multimodal sensing device for the test

In response to the current rapidly increasing market demand, the result of the project can provide information that each regional NMI can perform as a technical basis for the quantitative evaluation of acoustic cameras. Based on the results, it is expected that a calibration test can be conducted by comparing and verifying with more NMIs over an extended period of time in the future.

In-Jee Jung

injee@kriss.re.kr