Evaluation of Deep Neural Networks for Radar Based Point Cloud Classification
(2024) EITM01 20241Department of Electrical and Information Technology
- Abstract
- Thanks to the ever-increasing computational power, machine learning has become a staple in computer science. Many computer vision methods have become so reliable that their implementation in the surveillance industry is now the de facto standard for many object detection and classification tasks. The use of machine learning for surveillance is not confined to images and videos, however, it is also applicable to RADAR. Axis Communications produces several RADAR-based so- lutions, some use RADAR only, whereas others provide RADAR and video fusion technology. RADAR generates unordered sets of points, or point clouds. Point cloud classification is a fairly unexplored area of machine learning, in particular in the context of RADAR generated... (More)
- Thanks to the ever-increasing computational power, machine learning has become a staple in computer science. Many computer vision methods have become so reliable that their implementation in the surveillance industry is now the de facto standard for many object detection and classification tasks. The use of machine learning for surveillance is not confined to images and videos, however, it is also applicable to RADAR. Axis Communications produces several RADAR-based so- lutions, some use RADAR only, whereas others provide RADAR and video fusion technology. RADAR generates unordered sets of points, or point clouds. Point cloud classification is a fairly unexplored area of machine learning, in particular in the context of RADAR generated point clouds. In this thesis, we address the classification of moving clusters of RADAR point cloud data, which requires the classifier to consider several instances of an input cluster, where spatial as well as temporal information is present. We examine how two main classifier architectures perform on this data type and compare them to the existing classifier at Axis. Em- pirically, they show promising performance on the available data. However, the results also indicate that a more robust study might be required before employing the classifiers. (Less)
- Popular Abstract
- This study delves into a cutting-edge area: differentiating between humans and vehicles by combining radar devices with machine learning. Our research shows promising results but also highlights the need for more extensive testing to ensure reliability in real-world applications.
Radar surveillance offers the advantage of anonymity by allowing efficient mon- itoring without capturing personal details, ensuring privacy while still maintain- ing security. By integrating machine learning, radar systems become even more efficient and intelligent, capable of distinguishing between different objects and predicting potential threats with high accuracy. This advanced technology not only enhances safety and security but also respects individual... (More) - This study delves into a cutting-edge area: differentiating between humans and vehicles by combining radar devices with machine learning. Our research shows promising results but also highlights the need for more extensive testing to ensure reliability in real-world applications.
Radar surveillance offers the advantage of anonymity by allowing efficient mon- itoring without capturing personal details, ensuring privacy while still maintain- ing security. By integrating machine learning, radar systems become even more efficient and intelligent, capable of distinguishing between different objects and predicting potential threats with high accuracy. This advanced technology not only enhances safety and security but also respects individual privacy, providing a seamless and non-intrusive way to monitor and protect our world.
This thesis looks at how to distinguish objects over time using radar. Radar systems create groups of points, or "point clouds", which represent objects in their surroundings. Determining whether these point clouds are vehicles or humans using machine learning is still a relatively new field of research, especially when the data comes from radar. As these point clouds are followed over time, there is even more information than that provided by the radar to consider when tackling this task.
The study tests two main types of machine learning architectures on this data, both based on a technique known as PointNet. These are compared to another technique, developed by Axis Communications.
Our results show that these new techniques perform well on the available data, and provide novel insights into the possibilities of combining machine learning and radar. However, it also shows the shortcomings of these newer techniques, proposing further research opportunities that might find even better outcomes. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9167029
- author
- Sandelius, Alexander LU and Ronkainen, Emil LU
- supervisor
- organization
- course
- EITM01 20241
- year
- 2024
- type
- H2 - Master's Degree (Two Years)
- subject
- report number
- LU/LTH-EIT 2024-1001
- language
- English
- id
- 9167029
- date added to LUP
- 2024-06-26 13:26:39
- date last changed
- 2024-06-26 13:26:39
@misc{9167029, abstract = {{Thanks to the ever-increasing computational power, machine learning has become a staple in computer science. Many computer vision methods have become so reliable that their implementation in the surveillance industry is now the de facto standard for many object detection and classification tasks. The use of machine learning for surveillance is not confined to images and videos, however, it is also applicable to RADAR. Axis Communications produces several RADAR-based so- lutions, some use RADAR only, whereas others provide RADAR and video fusion technology. RADAR generates unordered sets of points, or point clouds. Point cloud classification is a fairly unexplored area of machine learning, in particular in the context of RADAR generated point clouds. In this thesis, we address the classification of moving clusters of RADAR point cloud data, which requires the classifier to consider several instances of an input cluster, where spatial as well as temporal information is present. We examine how two main classifier architectures perform on this data type and compare them to the existing classifier at Axis. Em- pirically, they show promising performance on the available data. However, the results also indicate that a more robust study might be required before employing the classifiers.}}, author = {{Sandelius, Alexander and Ronkainen, Emil}}, language = {{eng}}, note = {{Student Paper}}, title = {{Evaluation of Deep Neural Networks for Radar Based Point Cloud Classification}}, year = {{2024}}, }