A bird species identification method based on bird whistling

Abstract

The invention discloses a bird species identification method based on bird whistling, which comprises the following steps of: 1) performing short-time power spectrum transformation on bird whistling signals with noise in a training set, and detecting and screening out activity signals; 2) inputting the active signal into an external ear filter model and a middle ear filter model to obtain an enhanced auditory domain signal; Step 3) transforming the auditory domain signal into a critical band domain to obtain a time-critical band loudness spectrum; 4) extracting a linear prediction cepstrum coefficient of the loudness spectrum as a feature recognition quantity, and inputting the feature recognition quantity into a hidden Markov-deep neural network classifier for training to obtain a trainedbird buzzing classifier; and 5) extracting the linear prediction cepstrum coefficient of the loudness spectrum of the bird whistle to be identified as a feature identification quantity, and inputtingthe feature identification quantity into a trained bird whistle classifier to realize bird whistle identification. The method simulates the sound processing process of human ears, and solves the technical bottleneck of bird whistle automatic identification in a complex background noise environment.

Description

technical field [0001] The invention relates to the technical field of ornithology and remote automatic monitoring of birdsong; in particular, it relates to a method for identifying bird species based on birdsong. technical background [0002] Islands are important habitats for birds, and they are also important places for the protection and utilization of the ecological environment of birds. Developed countries such as the United States and the United Kingdom have established national-scale bird monitoring networks and achieved good results. my country has also carried out a large number of bird surveys and monitoring work, mainly focusing on the research on the species, quantity and distribution of terrestrial birds in a region. In order to promote the monitoring of birds in my country, in 2011, the Nanjing Institute of Environmental Science of the Ministry of Environmental Protection organized a demonstration monitoring of biodiversity (birds), and explored the construct...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problem of difficult identification and classification of bird species in the remote monitoring process such as unattended monitoring in complex wild environments such as islands / wetlands

Images