PhD Studentship: Three-dimensional object detection and segmentation in point clouds

at University of the West of England, Bristol
Published November 11, 2022
Location Bristol, United Kingdom
Category Machine Learning  
Job Type Scholarship  


Three-dimensional object detection and segmentation in point clouds in the context of aerospace CFD meshing

An opportunity to apply for a funded full-time PhD at UWE Bristol. The studentship will be jointly funded by UWE Bristol and Zenotech Ltd.

The CMV is located within the Bristol Robotics Laboratory (BRL), a joint venture between UWE Bristol and University of Bristol. The CMV specialises in real-world applications of computer vision and machine learning for the realisation of working prototypes and demonstrators, with a strong emphasis on 3D imaging and analysis, data capture, and modelling. This studentship will also receive support from the world-leading aerodynamic and software development teams at Zenotech, such as relevant datasets for project case study, and general domain expertise in Computational Fluid Dynamics (CFD) meshing.

Ref 2223-APR-CATE11

The expected start date of this studentship is 01 April 2023

The closing date for applications is 06 December 2022.

Studentship Details

Engineers and scientists are often faced with immense challenges when examining and processing complex 3D simulation data. This can be due to complexity and variability in the data representation in the 3D space, the potential large scale of the data itself, a high density of data points, and the inconvenience of data visualisation leading to unintuitive human perception of the 3D data. A typical example is the production of a Computational Fluid Dynamics (CFD) mesh based on aircraft geometry. A significant bottleneck in this activity is caused by the vast amount of time that CFD engineers spend on manually changing mesh rules and geometry to achieve an acceptable solution. Computer vision and machine learning, through enabling automatic detection and localisation of 3D objects in complex 3D point clouds, has tremendous potential for automating and improving the CFD meshing process. This study will investigate feature-based registration of point clouds and deep-learning-based 3D object detection in point clouds, in order to automatically select and tune mesh criteria for every type of aerodynamic device together with local contexts of an aircraft’s geometry.

For an informal discussion about the studentship, please email

Funding details

The studentship is available from 01 April 2023 for a period of 3.5 years, subject to satisfactory progress and includes a tax exempt stipend, which is currently £17, 668 per annum.

In addition, full-time tuition fees will be covered for up to 3.5 years (Home and Overseas).

Eligibility criteria

This will be a 3.5-year full-time commitment. The project is ideal for a self-motivated and enthusiastic student with a good honours degree (2:1 or equivalent, or above) in a relevant field, and evidence of further study at Masters level or equivalent. Please note, acceptance will also depend on evidence of readiness to pursue a research degree. Knowledge and experience of machine learning and coding (such as C++ or Python) is essential. Expertise in processing point cloud data is desirable.

How to apply

Please submit your application online. When prompted use the reference number 2223-APR-CATE11.

Supporting documentation: you will need to upload your research proposal, all of your degree certificates and transcripts, and a recognised English language qualification is required.

References: you will need to provide details of two referees as part of your application.

Closing date

The closing date for applications is 06 December 2022.

Further Information

It is expected that interviews will take place on weeks commencing 12 December 2022.