Microphone array steering with image-based source location

摘要

Methods and systems for beam forming an audio signal based on a location of an object relative to the listening device, the location being determined from positional data deduced from an optical image including the object. In an embodiment, an object's position is tracked based on video images of the object and the audio signal received from a microphone array located at a fixed position is filtered based on the tracked object position. Beam forming techniques may be applied to emphasize portions of an audio signal associated with sources near the object.

主权项

1. A method of focusing a microphone array, comprising:
receiving an audio signal from the microphone array, the microphone array disposed a distance away from a first component of an audio beam forming system; receiving a visual image from a plurality of images forming a video segment from an optical image sensor having a known position relative to the microphone array, the visual image including the first component within a field of view (FOV); determining positional information for the first component, relative to the microphone array by analyzing the visual image, and by determining a distance between the first component and the image sensor based on a predetermined size of the first component; tracking a change in the position of the first component relative to the image sensor based on changes in the visual image within the video segment; and filtering the audio signal to emphasize or de-emphasize audio sources proximate to the first component based on a beam forming algorithm that is a function of the positional information and on the change in the position of the first component.