Advancements in preprocessing, detection and classification techniques for ecoacoustic data: A comprehensive review for large-scale Passive Acoustic Monitoring

Expert Systems with Applications, Journal Year: 2024, Volume and Issue: 252, P. 124220 - 124220

Published: May 16, 2024

Computational ecoacoustics has seen significant growth in recent decades, facilitated by the reduced costs of digital sound recording devices and data storage. This progress enabled continuous monitoring vocal fauna through Passive Acoustic Monitoring (PAM), a technique used to record analyse environmental sounds study animal behaviours their habitats. While collection ecoacoustic become more accessible, effective analysis this information understand monitor populations remains major challenge. survey paper presents state-of-the-art approaches, with focus on applicability large-scale PAM. We emphasise importance PAM, as it enables extensive geographical coverage monitoring, crucial for comprehensive biodiversity assessment understanding ecological dynamics over wide areas diverse approach is particularly vital face rapid changes, provides insights into effects these changes broad array species ecosystems. As such, we outline most challenging tasks, including pre-processing, visualisation, labelling, detection, classification. Each evaluated according its strengths, weaknesses overall suitability recommendations are made future research directions.

Language: Английский

---

Utility of acoustic indices for ecological monitoring in complex sonic environments

Ecological Indicators, Journal Year: 2020, Volume and Issue: 121, P. 107114 - 107114

Published: Nov. 9, 2020

With the continued adoption of passive acoustic monitoring as a tool for rapid and high-resolution ecosystem monitoring, ecologists are increasingly making use suite indices to summarise sonic environment. Though these often reported well represent some aspect biology an ecosystem, degree which they confounded by various extraneous conditions is largely unknown. We conducted aural inventory across 23 field sites in Okinawa identify number unique animal sounds present recordings. Using values 'measured richness', we then examined how performance 11 commonly-used varied range (including presence absence insect stridulations, audible wind or rain, human-related sounds). Our analysis identified both well- poor-performing indices, those that were particularly sensitive conditions. Only two reflected measured richness full examined. A few relatively insensitive conditions, but no index correlated with when masked sound from broadband stridulating insects. results demonstrate considerable sensitivity most commonly used confounding highlighting challenges working large datasets collected field. make practical recommendations based on study design, aim identifying greatest utility indicators biodiversity management world's natural soundscapes.

Language: Английский

---

Ecoacoustics: acoustic sensing for biodiversity monitoring at scale

Remote Sensing in Ecology and Conservation, Journal Year: 2020, Volume and Issue: 6(3), P. 217 - 219

Published: Aug. 2, 2020

Nature sound recordings have been collected for over a hundred years, with an exponential increase since the 1950s (Ranft 2004). Most such were taken in order to describe and decipher animal communication. However, sounds of animals reveal more than behaviour: they also reflect structure functioning ecosystem which are part. The practice deploying remote acoustic sensors natural environments has systematized under term 'passive monitoring' (PAM), technical mostly used marine acoustics but then employed terrestrial aquatic (Gillespie et al. 2009; Marques 2012). Acoustic sensing distinct advantages make it complementary other sampling modalities. Like camera trapping, can be on land or water, all type habitats. An sensor advantage that capture wide spatial range (often 360° about 100 m habitats), is much less affected by occlusion imagery. It record continuously regularly long time period collect information full assemblage species as captures surrounding environnment. These properties ensure high effort rather low investment (Ciira wa Maina 2016; Hill 2018). In many studies, data analysed manually simple energy-based detectors, goal targeted monitoring single (Dawson Efford Gillespie Digby 2013). ambient those obtained automatic devices contain evidence list ecological information, as: absence/presence, population density, structure, community landscape architecture, phenology, reproduction period, migration interactions functions. Many these only become evident through large-scale analysis methods tailored data. Benefiting from growth recent decades scale processing, focus shift broader ecosystem-level questions, while using audio prime source evidence. This main ecoacoustics (Sueur Farina 2015). Ecoacoustic cover types environment deep sea tropical forest, biodiversity techniques LIDAR, satellite-based environmental DNA. Research ecoacoustic grown massively past 15 developing methodology hardware devices, signal machine learning visualization (see Sugai 2019, this issue, review). Particularly important move fundamental applied science, being deployed practical conservation Gordon 2019; Sertlek Znidersic 2020). Within context United Nations Sustainable Development Goals (UN SDGs), already demonstrated contribute useful evidence, complement sources. SDG 14 'Life below water' land', include threatened (Braune 2008; 2018), invasive (Grant Grant 2010), poaching (Hill noise pollution (both water) (Fairbrass 2018; 2019), degradation mountain ecosystems (Helbig-Bonitz Much progress at level classification, particular development indices one hand, use tools (Stowell Joly processing engineering work representations transformations 2014; Phillips At level, low-cost innovation challenge now connected so streams integrated (Roch 2017; Sethi 2018) systems able run analyses classification directly board. Large-scale should transferred application, widely management. tool. included large (i.e. national international) programmes, fashion standard design into programs. As cited above, there documented case studies methodological developments support move. We pleased introduce special issue Remote Sensing Ecology Conservation ecoacoustics, demonstrating across different value maturity methods. Methodologically, two broad paradigms reflected issue. One paradigm measures diversity soundscape computation indices: algorithmically straightforward highly scalable, yield implicit, holistic taxa. Sánchez-Giraldo (2020) Roca Van Opzeeland (2019) conduct very contexts – respectively forests Columbian Andes, underwater Southern Ocean quantify reliability indices. tackle encountered effect rain index computation, significant differences between Antarctic habitats set Campos-Cerqueira develop another extracting compressed long-term spectrogram representation. study first test efficiency policy, supporting idea research. second involves detecting counting individual events, often limited chosen target species. offers higher degree selectivity, performed approximately when automatically scale. thorough review passive techniques, propose good practices designing automated surveys. constitutes crucial step towards standardization collection. Smith define protocol suitable duration (multi-year) monitoring, focuses seasonally varying patterns peak activity. They demonstrate field produce comparable results manual transects, quarter effective survey effort. Yip measurements improve density estimates, serving proxy measure distance event autonomous recording unit. Both monophonic multi-channel recordings. either case, will usually multiple audible any given track. Lin Tsao provide roadmap source-separation including may help disentangle overlapping Sumitani interaction among vocalizing individuals characterized aid dimension reduction algorithm coupled new compact microphone array, leading localization. projects represented volume use, apply success principles geographic contexts. Altogether, contributions inform international policy.

Language: Английский

---

Acoustic animal identification using unsupervised learning

Methods in Ecology and Evolution, Journal Year: 2023, Volume and Issue: 14(6), P. 1500 - 1514

Published: April 17, 2023

Abstract Passive acoustic monitoring is usually presented as a complementary approach to wildlife communities and assessing ecosystem conditions. Automatic species detection methods support biodiversity analysis by providing information on the presence–absence of species, which allows understanding structure. Therefore, different alternatives have been proposed identify species. However, algorithms are parameterized specific Analysing multiple would help monitor quantify biodiversity, it includes taxonomic groups present in soundscape. We an unsupervised methodology for multi‐species call recognition from ecological soundscapes. The proposal based clustering algorithm, specifically learning algorithm multivariate data (LAMDA) 3pi automatically suggests number clusters associated with sonotypes. Emphasis was made improving segmentation audio analyse whole soundscape without parameterizing according each group. To estimate performance our proposal, we used four datasets locations, years habitats. These contain sounds major that dominate terrestrial soundscapes (birds, amphibians, mammals insects) audible ultrasonic spectra. presents performances between 75% 96% recognition. Using methodology, measured compared indices (ACI, NP, SO BI). Our performs assessments similar advantage about need prior knowledge recordings.

Language: Английский

---

Influence of recording devices and environmental noise on acoustic index scores: Implications for bird sound-based assessments

Ecological Indicators, Journal Year: 2024, Volume and Issue: 159, P. 111759 - 111759

Published: Feb. 1, 2024

Passive acoustic monitoring serves as a minimally invasive and effective method for biodiversity assessment, particularly in bird through the application of indices. However, use different recording devices types environmental noise (e.g., rain, wind, stream, traffic noise) lead to signal distortions that affect ecoacoustics Currently, there are no established guidelines specifying technical requirements signal-to-noise ratio (SNR) threshold accurate calculation To enhance accuracy indices assessments, this study investigated impact (rain, on In study, we selected six indices: Acoustic Complexity Index, Diversity Evenness Bioacoustic Entropy Normalized Difference Soundscape used four simultaneously record 104 h bird-sound data at same location. addition, 44 noisy signals with intensities were artificially synthesized comparison. The sound then analyze effects assessment. Our results showed (a) all affected by device used; (b) each index had sensitivities types; (c) was SNR above which effect negligible. This provides recommendations selection determines thresholds signals, contributing refinement protocols acquiring preprocessing These findings aim establish standardized acquisition future

Language: Английский

---

Soundscape Characterization Using Autoencoders and Unsupervised Learning

Sensors, Journal Year: 2024, Volume and Issue: 24(8), P. 2597 - 2597

Published: April 18, 2024

Passive acoustic monitoring (PAM) through recorder units (ARUs) shows promise in detecting early landscape changes linked to functional and structural patterns, including species richness, diversity, community interactions, human-induced threats. However, current approaches primarily rely on supervised methods, which require prior knowledge of collected datasets. This reliance poses challenges due the large volumes ARU data. In this work, we propose a non-supervised framework using autoencoders extract soundscape features. We applied dataset from Colombian landscapes captured by 31 audiomoth recorders. Our method generates clusters based autoencoder features represents cluster information with prototype spectrograms centroid decoder part neural network. analysis provides valuable insights into distribution temporal patterns various sound compositions within study area. By utilizing autoencoders, identify significant characterized recurring intense types across multiple frequency ranges. comprehensive understanding area's allows us pinpoint crucial sources gain deeper its environment. results encourage further exploration unsupervised algorithms as promising alternative path for environmental changes.

Language: Английский

---

Soundscape classification with convolutional neural networks reveals temporal and geographic patterns in ecoacoustic data

Ecological Indicators, Journal Year: 2022, Volume and Issue: 138, P. 108831 - 108831

Published: April 5, 2022

Interest in ecoacoustics has resulted an influx of acoustic data and novel methodologies to classify relate landscape sound activity biodiversity ecosystem health. However, indicators used summarize quantify the effects disturbances on can be inconsistent when applied across ecological gradients. This study dataset 487,148 min from 746 sites collected over 4 years Sonoma County, California, USA, by citizen scientists. We built a custom labeled soundscape components deep learning framework test our ability predict these components: human noise (Anthropophony), wildlife vocalizations (Biophony), weather phenomena (Geophony), Quiet periods, microphone Interference. These allowed us balance predicting variation environmental recordings relative time build dataset. patterns space that could useful for planning, conservation restoration, monitoring. describe pre-trained convolutional neural network, fine-tuned with reference data, classification achieving overall F0.75-score 0.88, precision 0.94, recall 0.80 five target components. deployed model all assess their hourly patterns. noted increase Biophony early morning evening, coinciding peak animal community vocalization (e.g., dawn chorus). Anthropophony increased during morning/daylight hours was lowest evenings, diurnal activity. Further, we examined related geographic properties at recording sites. decreased increasing distance major roads, while increased. were comparable more urban/developed agriculture/barren sites, significantly higher than less-developed shrubland, oak woodland, conifer forest results demonstrate broad is possible small datasets, classifications large gain knowledge.

Language: Английский

---

Distributed Acoustic Sensor Systems for Vehicle Detection and Classification

IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 31293 - 31303

Published: Jan. 1, 2023

Intelligent transport systems (ITS) are pivotal in the development of sustainable and green urban living. ITS is data-driven enabled by profusion sensors ranging from pneumatic tubes to smart cameras which used detect categorise passing vehicles. Simple sensors, such as tubes, successfully deployed for counting vehicles but not useful vehicle tracking or re-identification. Smart cameras, on other hand, collect comprehensive information suffer occlusion, patchy coverage, compromised vision adverse weather visibility. This work explores a novel data source based optical fibre acts uninterrupted length virtual using distributed acoustic sensor (DAS) system. Based real DAS collected field, we first present study latent features that uniquely identify given vehicle, otherwise referred signature. We formulate classification problem examines incoming extract signatures different types vehicle. To this end, implement methods comparative performance analysis reveals insights into potential role applications. pilot driven real-DAS validated promising results where type correctly identified with 94% accuracy size 95% accuracy.

Language: Английский

---

Land-use intensity and landscape structure drive the acoustic composition of grasslands

Agriculture Ecosystems & Environment, Journal Year: 2022, Volume and Issue: 328, P. 107845 - 107845

Published: Jan. 5, 2022

Language: Английский

---

Characterization of soundscapes with acoustic indices and clustering reveals phenology patterns in a subtropical rainforest

Ecological Indicators, Journal Year: 2025, Volume and Issue: 171, P. 113126 - 113126

Published: Jan. 27, 2025

Language: Английский

---