|
|
Radar Signal Processing and Fusion of Information
Reich, Bořek ; Maršík, Lukáš (referee) ; Zemčík, Pavel (advisor)
This bachelor's thesis focuses on fusion of millimetr-wave radar and camera. It proposes appropriate procedure and usage of these sensors for object detection. Object detection in this bachelor's thesis is focused on people and provides additional information about detected person. It proposes convolution neural network as means of person detection and fusion of mmWave radar and camera data. When person is detected, distance of person from sensors is found in mmWave radar point cloud. Testing is performed on input data from both sensors in different situations, in poorly lit, unkwonwn scenes, with unknown people etc. Distance measuring is validated with reference data.
|
|
|
Exploitation of Neural Networks for Fusion of Image and Non-Image Data
Reich, Bořek ; Maršík, Lukáš (referee) ; Zemčík, Pavel (advisor)
This master thesis uses convolutional neural networks to fuse image and non-image data. Both deep learning detection systems that rely only on image data (images from the camera) and that use both image and non-image data (images from the camera and data from the millimeter-wave radar) are studied in this thesis. A unique dataset containing raw millimeter-wave radar data and corresponding time-synchronized images from the camera was created for the purpose of comparing these two types of methods (data fusion methods and methods that utilize only image data). Furthermore, a time synchronization method for millimeter-wave radar and cameras using only off-the-shelf hardware is proposed. Finally, the created dataset is used to verify the detection capability of the system that uses only camera data and the fusion system that uses both millimeter-wave radar and camera data.
|
|
|
Object Tracking Using Radar
Rafajová, Kateřina ; Zemčík, Pavel (referee) ; Maršík, Lukáš (advisor)
The aim of this thesis is to modify the Gtrack algorithm from Texas Instruments so that it can be used in other projects in the Python programming language. The main advantage should be clear input of the algorithm parameters. Working with the algorithm through Texas Instruments is quite difficult, because the parameters affecting this algorithm are entered via the command line. Tuning the parameters is tedious and complicated. This diploma thesis should enable simpler testing of parameters on uploaded data. The Gtrack algorithm uses radar devices. The radar sends a signal which is reflected from objects around and is received back and processed. The result of the processing is a 2D/3D point cloud (space around the radar). From the information provided by the point cloud, it is possible to obtain data regarding objects in the vicinity of the radar. The basis is the Gtrack algorithm from Texas Instruments. However, this work allows it to be used independently of the Texas Instruments platform. The tracking algorithm recognizes individual objects and can monitor them for a certain period of time. Based on the recognition and tracking of the given object, it is possible to run for example some other event. The result is information about objects moving around the radar. The main benefit is the possibility of using this algorithm in other projects, the only condition is an input of the correctly formed point cloud.
|
|
|
Radar and Video Fusion
Galeta, Ondřej ; Reich, Bořek (referee) ; Maršík, Lukáš (advisor)
Hlavním cílem této práce je zvýšení kvality sledovaní objektů v třírozměrném prostoru pomocí fúze radarových a video dat získaných ze scén sledujících dopravu. Práce uvádí několik metod zabývajících se sbíráním párů bodů ze souřadnicových systémů obou senzorů pro prostorovou kalibraci s důrazem na automatizaci. Práce dále řeší problém asociace více detekovaných objektů naráz za pomoci modifikovaného Maďarského algoritmu. Práce rovněž ukazuje některé způsoby predikce vzdálenosti objektů detekovaných videem pomocí datasetu získaného z radarových dat. Výstup z fúzního modelu poskytuje prostorově rozšířené sledování dráhy a přesnější počítání sledovaných objektů než radarové a video modely samotné, což může být využito pro další analýzu dopravy v zkoumané oblasti.
|
|
|
Exploitation of Neural Networks for Fusion of Image and Non-Image Data
Reich, Bořek ; Maršík, Lukáš (referee) ; Zemčík, Pavel (advisor)
This master thesis uses convolutional neural networks to fuse image and non-image data. Both deep learning detection systems that rely only on image data (images from the camera) and that use both image and non-image data (images from the camera and data from the millimeter-wave radar) are studied in this thesis. A unique dataset containing raw millimeter-wave radar data and corresponding time-synchronized images from the camera was created for the purpose of comparing these two types of methods (data fusion methods and methods that utilize only image data). Furthermore, a time synchronization method for millimeter-wave radar and cameras using only off-the-shelf hardware is proposed. Finally, the created dataset is used to verify the detection capability of the system that uses only camera data and the fusion system that uses both millimeter-wave radar and camera data.
|
|
|
Radar Signal Processing and Fusion of Information
Reich, Bořek ; Maršík, Lukáš (referee) ; Zemčík, Pavel (advisor)
This bachelor's thesis focuses on fusion of millimetr-wave radar and camera. It proposes appropriate procedure and usage of these sensors for object detection. Object detection in this bachelor's thesis is focused on people and provides additional information about detected person. It proposes convolution neural network as means of person detection and fusion of mmWave radar and camera data. When person is detected, distance of person from sensors is found in mmWave radar point cloud. Testing is performed on input data from both sensors in different situations, in poorly lit, unkwonwn scenes, with unknown people etc. Distance measuring is validated with reference data.
|
guest ::
