A Study Of Acoustic Index Disturbance Factors For Bird Diversity Assessment

Since birds have been considered as one of the indicator species of ecological environment health,bird diversity can reflect biodiversity properly.Acoustic indices can be used to assess bird diversity,which have been applied to bird surveys and monitoring by researchers,while it is found that acoustic indices are easily impacted by different interference factors.At present,most research of acoustic index are focused on how to assess the bird diversity by using one or more acoustic indices in different ecological environment accurately.However,little research pay their attention on interference factors.These interference factors may result in inaccurate calculation of acoustic indices,further affect the results of assessment on bird diversity.In this thesis,three important interference factors,which are signal processing durations,recording devices and environmental noises,have been analyzed by seven widely used acoustic indices in bird diversity assessment.The main contents of this thesis are as follows:Firstly,the effect of signal processing duration on acoustic indices is investigated.Using sound signals collected in Wanzuitou Wetland Park of Guangzhou University City,eleven kinds of signal processing durations between 30 seconds and 3600 seconds are selected to calculate the acoustic indices under two modes of interval recording and continuous recording,respectively.The calculation results aim to analyze the correlation between various acoustic indices in different signal durations,also the correlation between different signal durations in a single acoustic index.The results show that the correlation between multiple acoustic indices decreases with increasing signal processing duration,therefore it is not recommended to use long signal processing durations for the acoustic indices calculation.Meanwhile,for a single acoustic index,researchers can refer to the time cost and the results of the hierarchical clustering analysis to choose the appropriate signal processing duration according to the actual requirements.Secondly,the effect of recording devices on acoustic index is investigated.Four different recording devices are placed at the same location in Wanzuitou Wetland Park of Guangzhou University City to record bird songs simultaneously,and each acoustic index of each recording signal is calculated.Compared the results,it shows that the calculation results of each acoustic index on different recording devices are inconsistent.When measuring the acoustic parameters of each recording equipment in the anechoic room,it is found that there are obvious differences in directivity,gain and amplitude frequency response.Calculating each acoustic index after adjusting the amplitude of each recording signals,results show that the value of each acoustic index remains unchanged corresponding to different recording devices,which means that the device gain has no effect on acoustic indices.Calculating each acoustic index after equalizing recording signals by designing the amplitude frequency response equalizer of each recording device,results show that each acoustic index corresponding to different recording devices tend to be consistent,which means that the influence of amplitude frequency response on acoustic index can be reduced to some extent by equalization algorithm.Finally,the effect of environmental noises on acoustic index is investigated.The bird data of nine common birds in Guangzhou City are collected from bird-sound website Xeno-canto.The environmental noise data of rain,stream,wind,and traffic are collected from the sound website Freesound.Eleven signal-to-noise ratio(SNR)sound files are synthesized by four types of environmental noises and each bird sound files,then calculate seven acoustic indices of each sound files,to analyze the sensitivity of each acoustic index to environmental noises and SNR.