Bird Species Identification Research Articles

Bioacoustics Data Analysis – A Taxonomy, Survey and Open Challenges

Biodiversity monitoring has become a critical task for governments and ecological research agencies for reducing significant loss of animal species. Existing monitoring methods are time-intensive and techniques such as tagging are also invasive and may adversely affect animals. Bioacoustics based monitoring is becoming an increasingly prominent non-invasive method, involving the passive recording of animal sounds. Bioacoustics analysis can provide deep insights into key environmental integrity issues such as biodiversity, density of individuals and present or absence of species. However, analysing environmental recordings is not a trivial task. In last decade several researchers have tried to apply machine learning methods to automatically extract insights from these recordings. To help current researchers and identify research gaps, this paper aims to summarise and classify these works in the form of a taxonomy of the various bioacoustics applications and analysis approaches. We also present a comprehensive survey of bioacoustics data analysis approaches with an emphasis on bird species identification. The survey first identifies common processing steps to analyse bioacoustics data. As bioacoustics monitoring has grown, so does the volume of raw acoustic data that must be processed. Accordingly, this survey examines how bioacoustics analysis techniques can be scaled to work with big data. We conclude with a review of open challenges in the bioacoustics domain, such as multiple species recognition, call interference and automatic selection of detectors.

High Accuracy Individual Identification Model of Crested Ibis (Nipponia Nippon) Based on Autoencoder With Self-Attention

As the population and the distribution of Crested Ibis (Nipponia nippon) become larger, it is necessary to propose a highly efficient census method to estimate the population size of the Crested Ibis. Passive acoustic monitoring (PAM) has a very good prospect for the Crested Ibis monitoring. To realize the automatic census of the Crested Ibis with PAM, the automatic individual identification method based on the vocalization is the key technology. A novel individual identification model was proposed in this paper, which built the autoencoder based on LSTM to obtain the meaningful latent representation from the raw recording directly, further, embedded self-attention and putted forward a combined training mode to achieve distinctive latent representation. With this model, nine Crested Ibis individuals were identified accurately, the highest accuracy is 0.971, and the average accuracy reaches 0.958. As for other three species, Little owl (Athene noctua), Chiffchaff (Phylloscopus collybita) and Tree pipit (Anthus trivialis), the better performances were achieved than the existing method, which means the proposed model can provide an alternative method for the individual identification of other bird species.

Identification of bird species and their conservation status in Tuti Island, Sudan

Identification of bird species and their conservation status in Tuti Island, Sudan

Identification of bird species and their conservation status in Tuti Island, Sudan

Identification of bird species and their conservation status in Tuti Island, Sudan

Bird Species Recognition Using Unsupervised Modeling of Individual Vocalization Elements

This paper investigates acoustic modeling for recognition of bird species from audio field recordings. First, the acoustic scene is decomposed into isolated segments, corresponding to detected sinusoids. Each segment is represented by a sequence of the frequency and normalized magnitude values of the sinusoid. The temporal evolution of these features is modeled using hidden Markov models HMMs. A novel method for an unsupervised modeling of individual bird vocalization elements is proposed. The element models are initialized using HMM-based clustering and then further trained using an iterative maximum likelihood label re-assignment procedure. State duration modeling, performed in a post-recognition stage, is explored. Finally, we developed a hybrid deep neural network—hidden Markov model. The developed acoustic models are employed for bird species identification, detection of specific species, and recognition of multiple bird species vocalizing in a given recording. The detection system employs score normalization. Recognition of multiple bird species is performed based on maximizing the likelihood of a set of segments on a subset of bird species models, with penalization based on Bayesian information criterion applied. Experimental evaluations are performed on more than 37 h of sound field recordings, containing vocalizations of 48 bird species, plus more than 16 h of non-bird sound recordings. Using 3 s of the detected signal, the best system achieved: identification accuracy of 98.7%, detection with the equal error rate of 2.7%, and recognition accuracy of 97.3% and 95.4% when vocalizations of multiple bird species are present, with the number of bird species known and estimated, respectively.

Bird species identification using spectrogram and dissimilarity approach

In this work, we investigate bird species identification starting from audio record-ings on eight quite challenging subsets taken from the LifeClef 2015 bird task contest database, in which the number of classes ranges from 23 to 915. The classification was addressed using textural features taken from spectrogram im-ages and the dissimilarity framework. The rationale behind it is that by using dissimilarity the classification system is less sensitive to the increase in the number of classes. A comprehensive set of experiments confirms this hypothesis. Although they cannot be directly compared to other results already published because in this application domain the works, in general, are not developed ex-actly on the same dataset, they overcome the state-of-the-art when we consider the number of classes involved in similar works. In the hardest scenario, we obtained an identification rate of 71% considering 915 species. We hope the subsets proposed in this work will also make future benchmarking possible.

Local compressed convex spectral embedding for bird species identification.

This paper proposes a multi-layer alternating sparse-dense framework for bird species identification. The framework takes audio recordings of bird vocalizations and produces compressed convex spectral embeddings (CCSE). Temporal and frequency modulations in bird vocalizations are ensnared by concatenating frames of the spectrogram, resulting in a high dimensional and highly sparse super-frame-based representation. Random projections are then used to compress these super-frames. Class-specific archetypal analysis is employed on the compressed super-frames for acoustic modeling, obtaining the convex-sparse CCSE representation. This representation efficiently captures species-specific discriminative information. However, many bird species exhibit high intra-species variations in their vocalizations, making it hard to appropriately model the whole repertoire of vocalizations using only one dictionary of archetypes. To overcome this, each class is clustered using Gaussian mixture models (GMM), and for each cluster, one dictionary of archetypes is learned. To calculate CCSE for any compressed super-frame, one dictionary from each class is chosen using the responsibilities of individual GMM components. The CCSE obtained using this GMM-archetypal analysis framework is referred to as local CCSE. Experimental results corroborate that local CCSE either outperforms or exhibits comparable performances to existing methods including support vector machine powered by dynamic kernels and deep neural networks.

Bird species identification via transfer learning from music genres

Humans possess the ability to apply previously acquired knowledge to deal with novel problems quite efficiently. Transfer Learning is inspired by exactly that ability and has been proposed to handle cases where the available data come from diverse feature spaces and/or distributions. This paper proposes to transfer knowledge existing in music genre classification to identify bird species, motivated by the existing acoustic similarities. We propose a Transfer Learning framework exploiting the probability density distributions of ten different music genres for acquiring a degree of affinity between the bird species and each music genre. To this end, we exploit a feature space transformation based on Echo State Networks. The results reveal a consistent average improvement of 11.2% in the identification accuracy of ten European bird species.

Diversity and Temporal Variation of Avian Fauna in a Rice Growing Landscape in Kurunegala District, Sri Lanka

Rice fields are man-made semi-aquatic ecosystems maintained for the purpose of paddy cultivation. Five contrastingly different successive culture stages viz, preparatory stage, vegetative stage, ripening stage, harvesting stage and post-harvest stage occur along a single paddy culture cycle. Although it has been recorded that a wide variety of fauna are associated with these culture stages, there is a dearth of information on the temporal variation of avian fauna along these culture stages. Therefore, a study was carried out in a rice growing landscape at Katugampola, a rural village in Kurunegala District, to investigate if there is any temporal variation of avian diversity along the culture stages of a paddy culture cycle there. It was hypothesized that the physical and biological characteristics vary along the successive culture stages, and that each culture stage offers a predictable habitat for a distinct assemblage of birds. The sampling protocol of the study included in situ observation, identification and enumeration of different bird species using the fixed radius point count sampling method, and assessing the physical and biological parameters such as the percentage area covered by paddy plants, height of the paddy plants, number of people, number of off road vehicles, number of cattle, number of dogs, sound level, wind speed and the associated insect fauna at each culture stage using appropriate standard field sampling techniques. The study was commenced in May 2014 and lasted until September 2014. Altogether 1892 bird counts were made during the study period where the birds belonged into 19 different species and to 12 different families. Physical and biological parameters also varied significantly (p<0.05 Oneway ANOVA) along the paddy culture cycle. Distinct bird assemblages characteristic to each culture stage were recorded during the study. For example, large predatory birds such as Cattle Egret, Intermediate Egret and Little Egret became abundant during the preparatory stage while small seed eating birds such as sparrows became abundant during the ripening and mature stages. Further, large seed eating birds such as Green Imperial Pigeon and Rose-ringed Parakeet became abundant both during the vegetative stage and the post-harvest stage. The temporal variation of the above bird assemblages appeared to have a direct relationship with the variation of the biological and physical parameters along the paddy culture cycle at Katugampola rice field. Keywords: Avian diversity, Rice growing landscapes, Paddy culture cycle

Using non-negative matrix factorisation to facilitate efficient bird species richness surveys

This paper considers computer-assisted learning of sound spectra in environmental recordings to facilitate manual bird species identification. Today, a variety of automated methods have been successfully applied for acoustic recognition of specific bird species. These methods are more effective for single targeted species detection. For in-field recordings, however, simultaneous vocalisations and unknown species usually make such methods less effective.In this study, we propose a non-negative matrix factorisation based method to facilitate manual bird species identification from environmental recordings. First, distinct sound spectra are extracted from each audio clip by applying non-negative matrix factorisation and clustering techniques. Based on these distinct sound spectra, a greedy algorithm is then designed to sample audio clips. Each sampled audio clip maximises the number of new spectra. People who follow this sampled sequence of audio clips should be able to identify the most species given a fixed number of audio clips. The efficiency is validated with annotated bird species per minute provided by experienced ornithologists.

Cluster analysis, fuzzy sets, and fuzzy logic models in bird identification

&lt;p&gt;In our resent research (Osadchiy at al., 2016) we considered the mathematical model for the identifying of bird species according to the results of inaccurate field measurements. We used the total length of the bird, the wingspan, the wingbeat frequency, and the flight as the input factors of the model. Testing the model on a hypothetical case of identifying some target species, like Rook, Common raven, Mallard, White Stork, and Lapwing revealed that this model can be used for bird species identification with definite limitations. However, in previous model we applied the recognition algorithm that was based on the classical sections of mathematical statistics. The limitations of those model are obvious - it does not take into account many characteristics and behavioral features of birds that cannot be represented in numerical form, like diurnal activity pattern and flocking behavior. In this case the possibility of using the traditional sections of mathematical statistics is quite limited. The present study is devoted to the development of a mathematical method for the identifying of the bird species that based on cluster analysis with fuzzy logic and fuzzy sets which extends the possibilities of the algorithm that was previously proposed in our research.&lt;/p&gt;

Actual and perceived collision risk for bird strikes at Namibian airports

At Namibia’s two major airports, Hosea Kutako International and Eros (domestic), 117 bird strike collision incidents were recorded between 2006 and 2010. A risk assessment, which included a proposed risk weighting methodology, was conducted at Hosea Kutako and Eros airports, which estimated the probability of an accident/collision as well as the consequence of such a collision. The assessment included surveys of bird occurrence frequencies as well as pilot interviews. The results of the risk assessments were compared with actual bird strike incidences for each species, frequency of occurrence of birds and pilot perceptions of species risk, in order to find whether risk assessment and pilot perception are reliable measures of potential bird strike incidence. White-backed Vulture Gyps africanus and Helmeted Guineafowl Numida meleagris were the highest risk species at both airports. They were also, after Crowned Lapwing Vanellus coronatus, the species most often observed by pilots. Bird strike records showed that Crowned Lapwing and Helmeted Guineafowl were also the most frequently struck birds at both airports. The study illustrates how combining risk assessment, pilot perception and bird strike history can benefit bird strike minimisation plans at airports through the rapid identification of priority bird species.

Identification of Bird Species Used to Make a Viking Age Feather Pillow

A grave containing the remains of a wooden boat was discovered in 1934 under a low mound in a bog at Øksnes in the Vesterålen islands of northern Norway. The boat grave dates to the 10th century in the Viking Age, and grave goods placed in the boat include an iron axe, a cowhide in which the body was wrapped, and pillow remains consisting of feather stuffing and a wool textile cover. A microscopic analysis of the feathers from a subsample of the pillow fill identified three avian orders: Anseriformes (eider); Suliformes (cormorant), and Charadriiformes (unspecified gull). It was possible to make one species-level identification of Great Cormorant (&lt;em&gt;Phalacrocorax carbo&lt;/em&gt;) and to narrow the gull types to the “white-headed” gull group. The sample was composed of a nearly equal mix of downy and pennaceous feather types. Downy feathers from gulls (Laridae) composed the majority of the material in this sample. While it is reported that feathers and down (assumed to be eider) were used in the Late Iron Age, this is the first successful attempt to identify bird species used in these materials and suggests that avian species identifications should be explored in other such burial items to enhance our understanding of human-wildlife interactions throughout Norse history.

Evaluation of an unmanned rotorcraft to monitor wintering waterbirds and coastal habitats in British Columbia, Canada

The effective protection of coastal and estuarine habitats requires reliable monitoring information on their use by waterbirds, and the use of small unmanned aircraft systems (UAS) may provide access to these habitats without disturbing birds. We evaluated the use of a rotary-wing UAS with a high-end consumer camera to identify and count wintering waterbirds at two coastal sites in British Columbia, Canada, in January 2015, and to map mudflat and marsh habitats. Photos of shorebirds, waterfowl, and seabird species were taken at varying altitudes, and disturbance of birds appeared minimal when the UAS was flown at heights ≥61 m. A ground resolution of ~1 cm/pixel was needed to discern plumage characteristics necessary to reliably identify birds. For some duck species, identification of females relied on body size or close association with a nearby male. Photographs were also used to derive accurate counts of shorebirds. For diving birds, accurate counts from photographs will require information on the proportion of birds on the water surface. Orthomosaics of coastal habitats were constructed with sufficient detail to assess ecological and geomorphological features. The UAS can therefore assist with bird species identification, population assessment, and characterization of habitat types.

A continuous linear optimal transport approach for pattern analysis in image datasets

We present a new approach to facilitate the application of the optimal transport metric to pattern recognition on image databases. The method is based on a linearized version of the optimal transport metric, which provides a linear embedding for the images. Hence, it enables shape and appearance modeling using linear geometric analysis techniques in the embedded space. In contrast to previous work, we use Monge׳s formulation of the optimal transport problem, which allows for reasonably fast computation of the linearized optimal transport embedding for large images. We demonstrate the application of the method to recover and visualize meaningful variations in a supervised-learning setting on several image datasets, including chromatin distribution in the nuclei of cells, galaxy morphologies, facial expressions, and bird species identification. We show that the new approach allows for high-resolution construction of modes of variations and discrimination and can enhance classification accuracy in a variety of image discrimination problems.

DNA Barcoding Identifies Illegal Parrot Trade.

Illegal trade threatens the survival of many wild species, and molecular forensics can shed light on various questions raised during the investigation of cases of illegal trade. Among these questions is the identity of the species involved. Here we report a case of a man who was caught in a Brazilian airport trying to travel with 58 avian eggs. He claimed they were quail eggs, but authorities suspected they were from parrots. The embryos never hatched and it was not possible to identify them based on morphology. As 29% of parrot species are endangered, the identity of the species involved was important to establish a stronger criminal case. Thus, we identified the embryos' species based on the analyses of mitochondrial DNA sequences (cytochrome c oxidase subunit I gene [COI] and 16S ribosomal DNA). Embryonic COI sequences were compared with those deposited in BOLD (The Barcode of Life Data System) while their 16S sequences were compared with GenBank sequences. Clustering analysis based on neighbor-joining was also performed using parrot COI and 16S sequences deposited in BOLD and GenBank. The results, based on both genes, indicated that 57 embryos were parrots (Alipiopsitta xanthops, Ara ararauna, and the [Amazona aestiva/A. ochrocephala] complex), and 1 was an owl. This kind of data can help criminal investigations and to design species-specific anti-poaching strategies, and demonstrate how DNA sequence analysis in the identification of bird species is a powerful conservation tool.

ICOPE-15-1013 Optical bird detection and species identification for prevention of bird strikes in wind farms

ICOPE-15-1013 Optical bird detection and species identification for prevention of bird strikes in wind farms

Bird Species Identification Based on Timbre and Pitch Features of Their Vocalization

Although wild bird watching has been a popular leisure activity, it is often the case that people have no idea about what kind of bird species they see or hear. To help people learn to identify bird species, this study proposes an automatic bird sound identification system. Considering bird vocalization can be generally divided into two categories, namely call and song, the proposed system is built upon a two-stage identification framework. The first stage performs call/song classification. If an unknown sound clip is classified as a call, it is then handled by a call identifier in the second stage; otherwise, it is handled by a song identifier. Both identifiers use two acoustic features, timbre and pitch, to determine which of the bird species the sound clip belong to. In using timbre features, bird sounds are converted into Mel-Frequency Cepstral Coefficients and their first derivatives and then analyzed using Gaussian mixture models. In using pitch feature, we convert bird sounds into MIDI note sequences and then use bigram models to analyze the dynamic change information of the notes. Our experiments, conducted using a database including twenty common bird species in the Taipei urban area, show that the proposed system can achieve 90.4% accuracy.

Automatic bird sound detection in long real-field recordings: Applications and tools

The primary purpose for pursuing this research is to present a modular approach that enables reliable automatic bird species identification on the basis of their sound emissions in the field. A practical and complete computer-based framework is proposed to detect and time-stamp particular bird species in continuous real field recordings. Acoustic detection of avian sounds can be used for the automatized monitoring of multiple bird taxa and querying in long-term recordings for species of interest for researchers, conservation practitioners, and decision makers, such as environmental indicator taxa and threatened species. This work describes two novel procedures and offers an open modular framework that detects and time-stamps online calls and songs of target bird species and is fast enough to report results in reasonable time for non-processed field recordings of many thousands files and is generic enough to accommodate any species. The framework is evaluated on two large corpora of real field data, targeting the calls and songs of American Robin Turdus migratorius, a Northamerican oscine passerine (true songbird) and the Common Kingfisher Alcedo atthis, a non-passerine species with a wide distribution throughout Eurasia and North Africa. With the aim of promoting the widespread use of digital autonomous recording units (ARUs) and species recognition technologies the processing code and a large corpus of audio recordings is provided in order to enable other researchers to perform and assess comparative experiments.

J0540301 風車バードストライク防止に向けた野鳥の画像検出システムの構築([J054-03]再生可能エネルギー(3))

As the number of wind turbines is increasing rapidly in the world, those pose a serious problem which rare birds impact turbine blades. It is getting more important to collect data for objective assessment of current birdstrike's problem. The authors try to offer optical base diagnosis system which detects flying birds in wind farms. The system composes of two units bird detection and bird species identification. In detection part, frame-differential method is applied to extract only birds from moving objects. In classification part, we estimate flapping frequency of birds from consecutive images so that the length scale of the bird can be estimated from aero-biological correlation. The present system allows a single-view monitoring to measure the distance of flying birds from the camera without employing stereoscopic photographing.