Identification Vehicle Movement Detection in Forest Area using MFCC and KNN

摘要

Audio-based monitoring systems have proven to be very useful tools for detecting such cases like gunshot detection, noise assessment in urban areas etc. There are several existing acoustic event detection algorithm that helps in different applications for area monitoring/ security concerns. It might normally be the case that there is hand saw/ axe / tiger roaring / animal call / vehicle engine sound from these kinds of incidents. Acoustic sensing at timely and correctly detection of the events is the main objective of event detection system. Timber theft or animal killing are the cases which should be identify and taken action by the security officers in the forest and this is only possible by image or acoustic sound detection. Image processing becomes too costly and not possible for wide area coverage but acoustic sensors are cheaper and effective. Automatic recognition of sound events can be valuable for efficient situations examination of acoustic scene. In this paper we concentrate on the problem of detecting the unauthorized vehicle movement in forest area. The primary goal is the continuous acoustic surveillance for a particular natural scene for illegal human activities (trespassers, killing animals etc.) in order to promptly alert an authorized officer for taking the appropriate measures.