[visionlist] [PhD scholarship] Semantic Understanding of 3D Point Clouds, $38, 092 p.a.

Tat-Jun Chin tat-jun.chin at adelaide.edu.au
Sun Mar 29 20:43:26 -04 2020

The UAiPhD is an innovative 4 year program, which includes completion of a
6 month industry placement. As a PhD student, you will have the opportunity
to work on an industry problem, while being supported by a University of
Adelaide supervisor and an industry supervisor for the duration of the
program. The exciting project with Maptek Pty Ltd is highlighted below.

Project name:
Semantic Understanding of 3D Point Clouds

Project description:
State-of-the-art LiDAR (light ranging and detection) devices, such as those
designed and manufactured by Maptek, are capable of producing accurate and
high-resolution 3D measurements of their surrounding environment. This
makes them ideally suited for surveying tasks in various industrial and
scientific applications. In particular, the usage of LiDAR devices for
surveying is standard in the global mining industry.

Apart from accurate point-wise 3D measurements, the point clouds generated
by a LiDAR scan also contain much richer information. For example, the
identity and structure of various objects in the environment can be
discerned from the point clouds. In fact, the ability to capture such
meaningful information has cemented the key role of LiDAR in the perception
module of autonomous vehicles and intelligent robots.

In the mining industry, the rich information contained in LiDAR point
clouds could also be exploited for various applications. The key to unlock
the valuable information in the point clouds are algorithms that can
extract the semantic information (e.g., identity and structure of objects)
that lie in the point clouds. While many such techniques have been
investigated for popular domains such as autonomous vehicles and robotics,
less attention has been devoted to developing such capabilities in the
mining domain.

The overall aim of this project is to develop algorithms that can achieve
semantic understanding of 3D point clouds, based on the mining applications
listed above. Due to the uniqueness of the mining environment, existing
algorithms that were developed for, e.g., urban/street scenes, indoor home
environments, cannot be directly applied to the mining applications. The
project will investigate novel algorithms based on machine learning (e.g.,
deep neural networks) and geometric optimisation to achieve the goals.

Location and Details:
The PhD project and industry placement component will take place in
Adelaide. While completing the PhD, students will be within the School of
Computer Science at the University of Adelaide and will receive principal
supervision from Associate Professor Tat-Jun Chin, as well as industry
supervision/guidance from Maptek.

Scholarship and Funding:
Successful students will receive a scholarship of approximately $38,092 tax
free per annum. This consists of a University funded Research Training
Program Scholarship of $28,092 per annum (2020 rate), alongside an Industry
Partner Scholarship of $10,000 per annum from Maptek. In addition, $5,000
per annum will be provided for overall project/research support costs.

Application instructions:
contact Associate Professor Tat-Jun Chin (http://cs.adelaide.edu.au/~tjchin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://visionscience.com/pipermail/visionlist_visionscience.com/attachments/20200330/97169bfb/attachment.html>

More information about the visionlist mailing list