Animal sound recognition based on double feature of spectrogram in real environment

Abstract

In this paper, we propose an animal sound recognition method in various noise environments with different Signal-to-Noise Ratios (SNRs). In real world, the ability to automatically recognize a wide range of animal sounds can analyze the habits and distributions of animals, which makes it possible to effectively monitor and protect them. However, due to the existence of different environments and noises, the existing method is difficult to ensure the recognition accuracy of animal sound in low SNR condition. To address this problem, this paper proposes double feature, which consists of projection feature and local binary pattern variance (LBPV) feature, combined with random forests for animal sound recognition. In feature extraction, an operation of projecting is made on spectrogram to generate the projection feature. Meanwhile, LPBV feature is generated by means of accumulating the corresponding variances of all pixels for every uniform local binary pattern (ULBP) in the spectrogram. As the experimental results show, the proposed method can recognize a wide range of animal sounds and still remains a recognition rate over 80% even under 10dB SNR.