Estimation of Transfer Coefficients and Signals of Sound-to-Light Conversion Device Blinky Under Saturation
Kosuke Nishida, Natsuki Ueno, Yuma Kinoshita, Nobutaka Ono · 2022 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) · 2022
In this study, we propose a method for estimating the intensity of the light signal emitted by a sound-to-light conversion device called Blinky from the saturated video signal obtained by a camera. A Blinky is a compact device that converts acoustic information into the light signal, and distributed Blinkies are applicable to various acoustic sensing frameworks without wired or wireless communication by monitoring them with a single video camera. Here, it is desired to obtain the intensity of the output light signal of the Blinky from the observed video signal; however, these are generally different owing to several factors such as light attenuation, background light, and saturation of brightness in the camera. We first calibrate coefficients describing the relationship between these signals using a given reference signal and then estimate the unknown signal of the Blinky using the calibrated coefficients. In each step, we consider the saturation effect in the observed video signal, which is a main contribution of this study. Experimental evaluations were performed under simulated and real environments, and their results indicated that the proposed method achieved higher estimation accuracy than the conventional method without consideration of the saturation effect.