Gunshot Detection Using Deep Learning and Acoustic Signal Processing for Safety of Society
摘要
Safety of the society is crucial for maintaining peace and security in a country, especially during the night hours. The increasing threat of robbing has led to a growing demand for continuous monitoring. The large societies where police stations are far from houses necessitate the development of an automatic gunshot detection system that can perform the duty independently without human intervention. Establishing robust, reliable, and prompt detection systems for illegal activities is essential for rural border security also. Other than the security of society the entire country’s border is a site of numerous illegal operations, with a high number of people being shot with firearms and entering the country illegally. Trespassing is common, and intruders often dig tunnels beneath the borders to gain access. Acoustic signal processing is used to identify gunshots with signals obtained through sensitive microphones. This paper presents a deep learning algorithm to detect the gunshots. The signal analysis, features, and the methods are described in the paper. The results show the 96% accuracy and perform better as compared to existing algorithms.