A Software Platform for Noise Reduction in Sound Sensor Equipped Drones

Abstract

A flying drone provides multiple video capturing options for filming videos. Since a noise is generated by the propellers and rotors of a drone, the quality of sound in the recorded video is quite low. Large drones are used singly in missions while small ones are used in formations or swarms. The small drones are proving to be useful in civilian applications. These are effective with multiple drones. Consideration of small drones for the applications, such as group flight, entertainment, and signal emission lead to deployment of networked drones. To solve the noise problem and develop group display applications, a software platform for these issues in networked drones is proposed. Noise reduction combines the active noise control and spectral subtraction. In addition, drones form group displays for an entertainment and displaying application. We develop a small-scale testbed to measure the service quality of proposed platform. The experimental results show that the proposed noise reduction produces a speech signal with up to 67.5% similarity to the original signal. It outperforms active noise control and spectral subtraction with the similarities of 53.1% and 39.6%, respectively. We see that drone formation can form a group display to show messages effectively.