About the Project
An opportunity exists to apply for a funded full-time PhD studentship in the Bristol Robotics Laboratory over a 3.5 year duration, including support for conference attendance and project consumables.
The expected start date of this studentship is 1st June 2018.
The closing date for applications is 28th March 2018
About the Studentship
The technology provider HP Inc. (HP) and the world renowned Bristol Robotics Laboratory (BRL) are working together on an exciting new research project to develop novel sensors and algorithms aimed at realising low-cost technology for the capture of unique surface characteristics from differing substrate types (e.g. paper, ceramic, plastic etc.), mainly for the purpose of object recognition and inspection.
Using a combination of both 3D structural and 2D spectral signals, the aim is to generate rich description of objects by recovering and describing properties that naturally occur at the surface, or which can be introduced, in order to realise a unique object recognition signature. The characterisation of surface properties will be investigated at both macro and microscale levels. Estimating 3D surface texture and reflectance at microscale is a challenging problem with a number of unsolved issues, and, requires the capture of highly dense microstructures.
This project will focus on:
(1) Using a range of optical/computer vision techniques to capture and recover surface characteristics of 3D surfaces objects at macro and microscale.
(2) Investigating the simultaneous recovery of multispectral signals and 3D structure at the microscopic scale.
(3) Developing novel algorithms to derive translation and rotation invariant signatures of different categories of objects and surfaces.
(4) Developing novel machine learning approaches (use of deep learning concepts) to characterise objects and identify their inherent surface properties.
3D depth recovery approaches such as structured light, photometric stereo and binocular stereo reconstruction will be considered, along with alternative methods based on phase shift interferometry and/or focus variation. Likewise, approaches looking at the narrow band spectral shifting data, captured using a hyperspectral imaging device (or a custom based multispectral image capture system) will also be investigated. In addition, this project will also investigate approaches looking at using localised properties (surface reflectance, surface texture and 3D properties) of a surface at microscale in order to build a more generalised meta-surface structure characterisation, say using a Simultaneous Localisation and Mapping (SLAM) approach.
The aim will be to establish signatures which can be captured using low-cost technology, are unique at the item level, can be readily compared and which have strong statistical significance.
Experienced BRL research academics from the Centre for Machine Vision (CMV), will work together with HP’s senior design engineers to provide guidance throughout the research programme. The CMV has developed internationally leading research and expertise in innovative 3D surface recovery solutions applied to different areas, including security, medical, agri-tech, digital advertising and others. Augmented by HP’s specialised skills at developing low-cost devices, this research will explore a range of state-of-art optical/computer vision techniques developed at the CMV to generate innovative solutions for the recognition and inspection of objects.
The PhD will be based at the CMV in BRL, with periods also spent working with the HP team in Bristol, and will benefit from links with the existing CDT programme within the BRL.
For an informal discussion about the studentship, please email Professor Melvyn Smith: [Email Address Removed] or Dr Khemraj Emrith: [Email Address Removed]
Funding details
The studentship is available from 1 June 2018 for a period of 3.5 years, subject to satisfactory progress and includes a tax exempt stipend at the standard EPSRC rate, which is currently £14,553 per annum. In addition, full-time tuition fees will be covered for up to 3.5 years (Home/EU rates only).
Eligibility criteria
Applicants must have a good honours degree (2.1 or equivalent) in subjects such as electronic engineering, computer science, computer vision, physics, sensor science.
A recognised English language qualification is required.
How to apply
Download and complete the Graduate School studentship application form and send it directly to the UWE Bristol Graduate School. Please ensure you include the title of the research project you propose to undertake, and detail why you are interested in undertaking this PhD project and what relevant knowledge, experience and qualifications you would bring to the research. Please see the Graduate School studentship application guidance notes for further information about how to complete the application form.