• Staffordshire University Featured PhD Programmes
  • University of Cambridge Featured PhD Programmes
  • Aberdeen University Featured PhD Programmes
  • University of Tasmania Featured PhD Programmes
  • University of Pennsylvania Featured PhD Programmes
University of Tasmania Featured PhD Programmes
Imperial College London Featured PhD Programmes
University of Bristol Featured PhD Programmes
Peter MacCallum Cancer Centre Featured PhD Programmes
University of Tasmania Featured PhD Programmes

An automated system for detecting and classifying marine organisms on manmade structures

This project is no longer listed in the FindAPhD
database and may not be available.

Click here to search the FindAPhD database
for PhD studentship opportunities
  • Full or part time
    Prof Smith
    Dr Kenny
  • Application Deadline
    No more applications being accepted
  • Competition Funded PhD Project (European/UK Students Only)
    Competition Funded PhD Project (European/UK Students Only)

Project Description

This is an extract of the research project. Simply click on “Apply on-line” above for an instant access to the complete version.

This project will develop an autonomous robotic instrument platform for monitoring and surveying underwater habitats located on manmade structures such as dock walls, piers and outflow pipes. A tentative scheme is shown in Figure 1. A machine vision system will be developed to capture 3D images of a vertical transect of the habitat using pulsed white light illumination and ultraviolet illumination (3D fluorescence photography). For the white light illumination images, high shutter speeds will be used to eliminate motion blur. 3D image processing and pattern classification will be implemented to identify and classify the visual features of different marine organisms, identify different types of marine organisms and may also give an indication to any new or invasive organisms. Target regions of the images will be segmented (Camargo and Smith, 2009) and features that characterise the shape and appearance of the image will be extracted from the segmented region. These features will be used as inputs into a robust classifier. An appropriate machine learning method (e.g. a Support Vector Machine (SVM)) will be used to identify the best classification model. This process has already been successfully implemented to visually identify diseases in plants (Camargo and Smith, 2009). Identification of species will be made more efficient than the previous research by using 3D images, which will produce depth information and also by the use of fluorescence photography, which will reveal further information not available using visible light illumination. For example; Figure 2 shows two images of the same organisms (found living on Dock walls in Liverpool). The UV images highlight certain internal organs and external structures of organisms. These features can be used to aid in the classification of different organisms.

Development of the instrument platform, imaging system and software will run concurrently to reduce development stalls. A Gantt chart showing the timescale for development is shown in Figure 3. Initially, 2D images of settlement panels and stills from video transects of dock walls, which are already available, will be used for the software development. The instrument platform will be tested in large tanks under laboratory conditions and in simulated operational environments (e.g. swimming pool) before being trialled in a real operational environment such as a dock.

The technology developed during this project is expected to be at technology readiness level 7 (using NASA definition of TRL7), by the end of the studentship. This is feasible, as the hardware (base station, instrument platform and lighting system) will be constructed from ‘off the shelf’ components, which will combine to form a new system, which will be controlled by a custom controller (constructed from an ‘off the shelf’ microcontroller).

Close guidance concerning hardware development will be supplied, to the student, primarily by Dr John Kenny (NOC) with further support from the mechanical and electronic engineering teams from the Ocean Technology and Engineering Group (OTEG) at NOC Liverpool, which has considerable experience in the development and deployment of marine systems, as well as an extensive store of equipment relevant to this project. Additional financial support from OTEG Liverpool, will be provided to purchase components and materials for the construction of the instrument platform. The software will be developed from existing image processing, pattern classification and machine learning method algorithms and processes, of which, the primary supervisor (Professor Jeremy Smith) is a leading authority. Guidance on the biological aspects of the project, including generation of source (learning) data for the machine learning algorithms, will be provided by Dr Matthew Spencer.

When complete, the system and processes developed throughout the course of this project can be packaged into a methodology, which can be used to form the basis of a standardised monitoring tool for evaluating the health of marine habitats on manmade structures (dock walls etc). A standardised monitoring tool for evaluating the health of marine habitats on manmade structures has the potential for widespread adoption nationally and internationally.

Funding Notes

Competitive tuition fee, research costs and stipend (£14,056 tax free) from the NERC Doctoral Training Partnership “Understanding the Earth, Atmosphere and Ocean” (DTP website: http://www.liv.ac.uk/studentships-earth-atmosphere-ocean/) led by the University of Liverpool, the National Oceanographic Centre and the University of Manchester. The studentship is granted for a period of 42 months. Further details on eligibility, how to apply, deadlines for applications and interview dates can be found on the website. EU students are eligible for a fee-only award.

References

Camargo, A. and Smith, J.S. (2008). An image-processing based algorithm to automatically identify plant diseas visual symptoms. Biosystems engineering. 102(2008) pp. 9-21

Camargo, A. and Smith, J.S. (2009). Image pattern classification for the identification of disease causing agents in plants. Computers and electronics in agriculture. 66(2009) pp. 121-125

Share this page:

Cookie Policy    X