Robotics and Semantic Systems

Computer Science | LTH | Lund University

Denna sida på svenska This page in English


CS MSc Thesis Presentations 16 January 2023


From: 2023-01-16 14:15 to 16:00
Place: E:4130 (Lucas)
Contact: birger [dot] swahn [at] cs [dot] lth [dot] se
Save event to your calendar

Two Computer Science MSc theses to be presented on 16 January

Monday, 16 January there will be two master thesis presentations in Computer Science at Lund University, Faculty of Engineering.

The presentations will take place in room E:4130 (Lucas).

Note to potential opponents: Register as an opponent to the presentation of your choice by sending an email to the examiner for that presentation ( Do not forget to specify the presentation you register for! Note that the number of opponents may be limited (often to two), so you might be forced to choose another presentation if you register too late. Registrations are individual, just as the oppositions are! More instructions are found on this page.

14:15-15:00 in E:4130 (Lucas)

Presenters: Johan Lindberg, Amanda Zarkout
Title: Evaluation of Spot as an instrument to provide autonomous data collection
Examiner: Elin Anna Topp
Supervisors: Volker Krueger (LTH), Mathias Haage (Cognibotics AB)

Improved digitization on construction sites results in more efficient feedback for day-to-day work. This project evaluates data gathering and data analysis methods based on the inspection robot Spot from Boston Dynamics. The robot has during this Autumn collected data each week on the construction site Vipan in Lund. This is the material used in this thesis. For this work to program modules for Spot have been developed. They allow data gathering and analysis, both online and offline, on construction sites. The availability of Swedish construction site datasets has been explored. Very little such material is publicly available. The performance of two other public datasets has therefore been explored, as well as the work of analyzing collected material. A method of using small annotated datasets for the analysis of specific tasks has been evaluated throughout the work. 

Link to popular science summary: To be uploaded

15:15-16:00 in E:4130 (Lucas)

Presenters: Philip Afsén, Kasper Boye Frick
Title: Aiding in the visualization of tagged items in a point cloud
Examiner: Elin Anna Topp
Supervisors: Anders Heyden (LTH), Niklas Hansson (AXIS), Carl-Axel Alm (AXIS)

LiDAR point clouds are often incomplete due to single viewpoint, or blockage and are therefore hard to interpret. In this paper, we are evaluating different ways of visualizing annotated data in a point cloud and how to implement the different methods. We then present our contributions. First, we evaluate a ground mesh based on the Poisson Surface Reconstruction method and Ball Pivoting Algorithm. Visualizing cars was done by predicting their shape with the neural network PoinTr. We came to the conclusion that pedestrian point clouds are too sparse to be able to visualize with the methods available today. The last thing to be evaluated was trees which were visualized with a rotation method combined with a mesh. Our results show the difference made by visualizing objects in the point cloud. We conclude the report with a discussion about the results and future directions for work in this area.

Link to popular science summary: To be uploaded