A New Approach to Real Time Impulsive Sound Detection for Surveillance Applications
This work addresses audio surveillance for public safety, but it is incremental as it adapts existing methods rather than introducing new ones.
The paper reviews impulsive sound detection algorithms for surveillance and adapts Warped Linear Prediction (WLP) from impulsive noise detection, showing it can be used for detecting dangerous sounds like gunshots or explosions.
Most of the surveillance systems for public safety are solely based on one or more video cameras. These camera systems have some drawbacks such that they have poor performance in adverse weather conditions or during night time. Therefore most of the time, some other sensors should accompany to video cameras. Although audio surveillance is in its early stage, there has been considerable amount of work in this area in the last decade. In this paper we make a review of impulsive sound detection algorithms. Sounds from dangerous events such as gunshots, explosions, human screaming can be classified as impulsive sounds, so this paper reviews all impulsive sound detection algorithms along with impulsive noise detection algorithms although they progress in their own path. These dangerous sound events have no other detection means except audio. We try to adapt some algorithms used in impulsive noise detection to the area of impulsive sound detection. Tests show that Warped Linear Prediction (WLP) can be used for impulsive sound detection.