Difference between revisions of "Projects:2019s1-151 Raspberry Pi as a Core Device for Efficient Biological Field Survey Data Collection"

From Projects
Jump to: navigation, search
(Introduction)
 
(One intermediate revision by the same user not shown)
Line 2: Line 2:
 
== Abstract ==
 
== Abstract ==
 
Biological surveys rely increasingly on electronic devices for acoustic data collection, as well as newer methods for processing and filtering enormous raw datasets. The increasing availability of inexpensive but professional-grade electronic components provides the potential for the creation of customised and user-updated devices suitable for a wider range of applications, reducing the reliance of biologists on proprietary devices and upgrade timelines. However, creating reliable devices from multiple electronic parts that function as desired can be a non-trivial exercise that is beyond the training of many biologists.
 
Biological surveys rely increasingly on electronic devices for acoustic data collection, as well as newer methods for processing and filtering enormous raw datasets. The increasing availability of inexpensive but professional-grade electronic components provides the potential for the creation of customised and user-updated devices suitable for a wider range of applications, reducing the reliance of biologists on proprietary devices and upgrade timelines. However, creating reliable devices from multiple electronic parts that function as desired can be a non-trivial exercise that is beyond the training of many biologists.
The purpose of this project was to build a customised and flexible device for the long-term collection of large-scale ultrasonic recordings relevant to bat identifications in biodiversity surveys. The device also needs to be capable of on-board processing of big data and sending small subsets of data over the mobile network for remote verification of target detections. Further, this thesis summarises how straightforward a build process can be, where the technical difficulties lie and tests the ability of the device for onboard processing of large datasets
+
The purpose of this project was to build a customised and flexible device for the long-term collection of large-scale ultrasonic recordings relevant to bat identifications in biodiversity surveys. The device also needs to be capable of on-board processing of big data and sending small subsets of data over the mobile network for remote verification of target detections. Further, this thesis summarises how straightforward a build process can be, where the technical difficulties lie and test the ability of the device for onboard processing of large datasets
 +
 
 +
== Introduction ==
 +
Use of acoustic devices such as Ecological Acoustic Recorders (EAR) has revolutionized understanding of terrestrial animal species and aquatic life. Acoustic recordings have enabled researchers to gather information which has also helped in the understanding of the evolution of animal species.
 +
Acoustic recorders are used to detect sounds produced by various animal species such as birds, frogs, bats, and cetaceans (dolphins and whales). Monitoring of organisms in remote habitats has mainly been done using acoustic instruments. For instance, monitoring of animal species in forests to understand their interactions requires observation effort and resources, since some of the forests are found in very remote areas. Personnel undertaking such studies may face difficulties due to harsh environments, which may have some species present that are listed as threatened under environmental legislation. The information from such locations is essential in the understanding of the ecosystem, which helps in the conservation of the endangered species, and particularly for environmental impact assessments for development projects. For the biological surveys carried out in remote locations, acoustic recorders become useful in the recording of biologically produced sounds, which are then analysed for the generation of data used to identify different species, their population sizes and even for monitoring their migration.
 +
Acoustic devices collect data which is then used to identify and quantify the activity of species. In Automated Remote Biodiversity Monitoring Network (ARBIMON), hardware and software devices are combined for automation of data, recording and management of data that enables identification of the species based on the data recordings. Apart from marine life, acoustic recordings have proved essential in biological surveys involving birds, frogs, bats, and cetaceans. Use of acoustic devices in the study of bats has helped to save on the amount of time and resources used in the traditional study of the bats and even birds. As compared to simple visual or capture-based observation in the field, acoustic devices have proved to be relatively accurate and more reliable, verifiable and expedient. These acoustic surveys involve the study of sounds produced by various animal species and even reception and dispersion of the sounds by the animal species. In this project developed a professional-grade device that can record the bat calls using an Ultrasound Microphone, with pre-processed recorded data classified on-board, and outcomes transmitted over the mobile network for future validation and identification of species.
 +
 
 +
== Motivation ==
 +
Use of acoustic methods to study animal species was first used by Ivan Regen, who was a scientist from Slovenia who used insects to study bioacoustics. Investigation of animal species is performed using acoustic recording devices. There are a variety of commercially produced sound recorders on the market, some of which allow storage of sound recordings on-board for later processing. Ultrasonic sound recorders are used for very high-frequency sounds, such as bat echolocation. Other devices are suitable for sounds in the human-audible range, and in water. Computers are then used to analyse the recorded sounds, which provide information useful in biological research on the various animal species, but especially they allow the identification of species. Unlike the traditional capture-based field surveys, use of acoustic recording increases the encounter rate of species.
 +
One of the methods used in acoustic surveys is the use of Passive Acoustic Monitoring (PAM), which enables remote monitoring of terrestrial animal species. The technique makes use of Autonomic Recording Units (ARU), which are devices that are left in the field to detect sounds and enable researchers to identify vocalizing species. Use of PAM has enabled improved detection of rare species and it has been possible to document biodiversity existing in localities. Some of the species are very sensitive to trapping or intrusion of their natural, and they can only be studied using the acoustic recordings. That way, the data collected is more representative because the animals continue natural behaviour without any human interference. Use of acoustic recordings is also useful where the terrain restricts access by investigators. Some of the acoustic devices are used together with video and photographic devices to enhance the accuracy of the distribution and abundance estimates. Some simple qualitative surveys use hydrophones to detect the type of species present in a given region. These surveys are carried out when there is little information on, for example, cetacean population sizes, and the researchers are only interested in determining the species that are found in the locality.
 +
Using standardized equipment and survey methods can detect changes in the population size and even migration of various species. Acoustics-based research is carried out when it is impossible to use other methods of survey. For instance, researchers have used acoustic recording devices to investigate frogs and ensure their biodiversity is maintained in efforts to ensure that they remain protected, and similar efforts are regularly applied to many vocalising species in environmental impact assessments. Popular acoustic devices that are used in biological surveys include the Anabat Swift Passive Bat Detector by Titley Scientific and Song Meter-4(SM4 Bat),. These can produce full-spectrum recordings and are robust to the environmental conditions. The Anabat Swift operates within a frequency range of 0-250kHZ(depending on the type of microphone), and the SM4 has inbuilt microphones which are used to record sounds from 20HZ- 48kHz. These systems are relatively costly (Anabat Swift Bat detector excluding microphone cost $1299, Song Meter-4 Bat Ultrasound Recorder excluding microphone costs $1300), and the recorded data is stored on flash memory such as SD cards. The devices allow the researchers to set a continuous monitoring or scheduled monitoring of the animal species.
 +
The study of the different ecological systems is vital in the translocation of animal species. The survival of any species in a new environment depends on its behaviour before the translocation. The information of the species in its natural habitat is essential when taking them to a new environment. The behaviour of the animal species is studied by using collar-mounted sensors which record data which is then useful in conservation of endangered species.
 +
 
 +
== Objectives ==
 +
The overall aim of the project was to create a new flexible device that could function as an acoustic sentinel for detecting infrequent encounter events of rare or invasive species. There were various requirements for the device:
 +
 
 +
  - The device should be able to connect all the sensors (Ultrasound Microphone) and peripherals (Quectel LTE Standard EC25, Compute Module) successfully.
 +
  - The device should access the sound files recorded by a commercial high-quality bioacoustics’ recorder (Pettersson Elektronik M500-384).
 +
  - The device should conduct some level of on-board processing and data compression on these recorded sound files.
 +
  - The device should send the positive identifications of a target signal over a network to a remote investigator for their verification.
 +
  - The device should be capable of storing bulk data for later verification processes.
 +
  - The device should interface over a network to allow remote software upgrades
  
 
== Team Members ==
 
== Team Members ==
Line 10: Line 31:
  
 
== Supervisors ==
 
== Supervisors ==
: Prof LangFord White
+
: Prof Langford White
 
: Dr Kyle Armstrong
 
: Dr Kyle Armstrong
 
 
== Objectivies ==
 
:
 

Latest revision as of 15:59, 6 November 2019

Abstract

Biological surveys rely increasingly on electronic devices for acoustic data collection, as well as newer methods for processing and filtering enormous raw datasets. The increasing availability of inexpensive but professional-grade electronic components provides the potential for the creation of customised and user-updated devices suitable for a wider range of applications, reducing the reliance of biologists on proprietary devices and upgrade timelines. However, creating reliable devices from multiple electronic parts that function as desired can be a non-trivial exercise that is beyond the training of many biologists. The purpose of this project was to build a customised and flexible device for the long-term collection of large-scale ultrasonic recordings relevant to bat identifications in biodiversity surveys. The device also needs to be capable of on-board processing of big data and sending small subsets of data over the mobile network for remote verification of target detections. Further, this thesis summarises how straightforward a build process can be, where the technical difficulties lie and test the ability of the device for onboard processing of large datasets

Introduction

Use of acoustic devices such as Ecological Acoustic Recorders (EAR) has revolutionized understanding of terrestrial animal species and aquatic life. Acoustic recordings have enabled researchers to gather information which has also helped in the understanding of the evolution of animal species. Acoustic recorders are used to detect sounds produced by various animal species such as birds, frogs, bats, and cetaceans (dolphins and whales). Monitoring of organisms in remote habitats has mainly been done using acoustic instruments. For instance, monitoring of animal species in forests to understand their interactions requires observation effort and resources, since some of the forests are found in very remote areas. Personnel undertaking such studies may face difficulties due to harsh environments, which may have some species present that are listed as threatened under environmental legislation. The information from such locations is essential in the understanding of the ecosystem, which helps in the conservation of the endangered species, and particularly for environmental impact assessments for development projects. For the biological surveys carried out in remote locations, acoustic recorders become useful in the recording of biologically produced sounds, which are then analysed for the generation of data used to identify different species, their population sizes and even for monitoring their migration. Acoustic devices collect data which is then used to identify and quantify the activity of species. In Automated Remote Biodiversity Monitoring Network (ARBIMON), hardware and software devices are combined for automation of data, recording and management of data that enables identification of the species based on the data recordings. Apart from marine life, acoustic recordings have proved essential in biological surveys involving birds, frogs, bats, and cetaceans. Use of acoustic devices in the study of bats has helped to save on the amount of time and resources used in the traditional study of the bats and even birds. As compared to simple visual or capture-based observation in the field, acoustic devices have proved to be relatively accurate and more reliable, verifiable and expedient. These acoustic surveys involve the study of sounds produced by various animal species and even reception and dispersion of the sounds by the animal species. In this project developed a professional-grade device that can record the bat calls using an Ultrasound Microphone, with pre-processed recorded data classified on-board, and outcomes transmitted over the mobile network for future validation and identification of species.

Motivation

Use of acoustic methods to study animal species was first used by Ivan Regen, who was a scientist from Slovenia who used insects to study bioacoustics. Investigation of animal species is performed using acoustic recording devices. There are a variety of commercially produced sound recorders on the market, some of which allow storage of sound recordings on-board for later processing. Ultrasonic sound recorders are used for very high-frequency sounds, such as bat echolocation. Other devices are suitable for sounds in the human-audible range, and in water. Computers are then used to analyse the recorded sounds, which provide information useful in biological research on the various animal species, but especially they allow the identification of species. Unlike the traditional capture-based field surveys, use of acoustic recording increases the encounter rate of species. One of the methods used in acoustic surveys is the use of Passive Acoustic Monitoring (PAM), which enables remote monitoring of terrestrial animal species. The technique makes use of Autonomic Recording Units (ARU), which are devices that are left in the field to detect sounds and enable researchers to identify vocalizing species. Use of PAM has enabled improved detection of rare species and it has been possible to document biodiversity existing in localities. Some of the species are very sensitive to trapping or intrusion of their natural, and they can only be studied using the acoustic recordings. That way, the data collected is more representative because the animals continue natural behaviour without any human interference. Use of acoustic recordings is also useful where the terrain restricts access by investigators. Some of the acoustic devices are used together with video and photographic devices to enhance the accuracy of the distribution and abundance estimates. Some simple qualitative surveys use hydrophones to detect the type of species present in a given region. These surveys are carried out when there is little information on, for example, cetacean population sizes, and the researchers are only interested in determining the species that are found in the locality. Using standardized equipment and survey methods can detect changes in the population size and even migration of various species. Acoustics-based research is carried out when it is impossible to use other methods of survey. For instance, researchers have used acoustic recording devices to investigate frogs and ensure their biodiversity is maintained in efforts to ensure that they remain protected, and similar efforts are regularly applied to many vocalising species in environmental impact assessments. Popular acoustic devices that are used in biological surveys include the Anabat Swift Passive Bat Detector by Titley Scientific and Song Meter-4(SM4 Bat),. These can produce full-spectrum recordings and are robust to the environmental conditions. The Anabat Swift operates within a frequency range of 0-250kHZ(depending on the type of microphone), and the SM4 has inbuilt microphones which are used to record sounds from 20HZ- 48kHz. These systems are relatively costly (Anabat Swift Bat detector excluding microphone cost $1299, Song Meter-4 Bat Ultrasound Recorder excluding microphone costs $1300), and the recorded data is stored on flash memory such as SD cards. The devices allow the researchers to set a continuous monitoring or scheduled monitoring of the animal species. The study of the different ecological systems is vital in the translocation of animal species. The survival of any species in a new environment depends on its behaviour before the translocation. The information of the species in its natural habitat is essential when taking them to a new environment. The behaviour of the animal species is studied by using collar-mounted sensors which record data which is then useful in conservation of endangered species.

Objectives

The overall aim of the project was to create a new flexible device that could function as an acoustic sentinel for detecting infrequent encounter events of rare or invasive species. There were various requirements for the device:

  - The device should be able to connect all the sensors (Ultrasound Microphone) and peripherals (Quectel LTE Standard EC25, Compute Module) successfully.
  - The device should access the sound files recorded by a commercial high-quality bioacoustics’ recorder (Pettersson Elektronik M500-384).
  - The device should conduct some level of on-board processing and data compression on these recorded sound files.
  - The device should send the positive identifications of a target signal over a network to a remote investigator for their verification.
  - The device should be capable of storing bulk data for later verification processes.
  - The device should interface over a network to allow remote software upgrades

Team Members

Ganesh Methra.
Karthik Allipur.


Supervisors

Prof Langford White
Dr Kyle Armstrong