|
Exploitation of Neural Networks for Fusion of Image and Non-Image Data
Pařilová, Michaela ; Reich, Bořek (referee) ; Zemčík, Pavel (advisor)
This bachelor's thesis deals with exploitation of neural networks for fusion of image and non-image data. It primarily focuses on comparing detection algorithms for image data (data obtained from a camera) with detection algorithms using the fusion of image data and non-image data (data obtained from a camera and radar). As part of the comparison of these two approaches, an own fusion model was created, which was subsequently tested on the selected data set.
|
|
Forest monitoring using the GEDI sensor
Šedová, Adéla ; Potůčková, Markéta (advisor) ; Moudrý, Vítězslav (referee)
The overall objective of this thesis was to explore the use of GEDI and its integration with Airborne Laser Scanning (ALS) for large scale forest monitoring. The study was carried out using a sample of GEDI footprints that fell into the timeline of three available ALS datasets that were acquired during the same year. The study area, located in Southeast of Czechia, is covered with mature 121-year-old forest monoculture of Norway spruce (Picea abies), and due to frequent disturbances caused by infestation is closely monitored as a part of research on forest dynamic. As a result, the forest is highly fragmented, and due to its dynamic character, close dates of acquisitions were preferred to a larger dataset. Canopy gaps and low tree densities are known to pose a challenge for large-footprint full-waveform LiDARs. The specific of GEDI sensor, such as its footprint size, were specially designed to overcome these challenges. The options of optimising GEDI's geolocation accuracy were explored. A tool for integrating GEDI and ALS data, the GEDI Simulator, was used to standardise both data sources and derive elevation height, Relative Height (RH) and Canopy Cover Fraction (CCF). The metrics were derived from real GEDI waveforms, simulated GEDI-like waveforms, and calculated from the ALS point cloud, and...
|
|
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.
|
|
Automatic detection of driving lanes geometry based on aerial images and existing spatial data
Růžička, Jakub
The aim of the thesis is to develop a method to identify driving lanes based on aerial images and existing spatial data. The proposed method uses up to date available data in which it identifies road surface marking (RSM). Polygons classified as RSM are further processed to obtain their vector line representation as the first partial result. While processing RSM vectors further, borders of driving lanes are modelled as the second partial result. Furthermore, attempts were done to be able to automatically distinguish between solid and broken lines for a higher amount of information contained in the resulting dataset. Proposed algorithms were tested in 20 case study areas and results are presented further in this thesis. The overall correctness as well as the positional accuracy proves effectivity of the method. However, several shortcomings were identified and are discussed as well as possible solutions for them are suggested. The text is accompanied by more than 70 figures to offer a clear perspective on the topic. The thesis is organised as follows: First, Introduction and Literature review are presented including the problem background, author's motivation, state of the art and contribution of the thesis. Secondly, technical and legal requirements of RSM are presented as well as theoretical concepts and...
|
|
Multimodal System for Multi-Object Tracking in Real-Time
Kučera, Adam ; Šátek, Václav (referee) ; Rozman, Jaroslav (advisor)
This thesis deals with the topic of multi-object multi-sensor tracking. A conventional track-oriented multiple hypothesis tracking (TOMHT) pipeline is implemented in C++ programming language and an implementable interface is designed, enabling to easily extend the core algorithm with arbitrary sensors and measured target attributes, making the system multimodal, i.e.\ applicable in heterogeneous systems of sensors. A novel algorithm for solving combinatorial optimization arising in TOMHT is proposed. Finally, few example implementations of the interface are provided and the system is evaluated in simulated and real-world scenarios.
|
|
Linked Data Integration
Michelfeit, Jan ; Knap, Tomáš (advisor) ; Klímek, Jakub (referee)
Linked Data have emerged as a successful publication format which could mean to structured data what Web meant to documents. The strength of Linked Data is in its fitness for integration of data from multiple sources. Linked Data integration opens door to new opportunities but also poses new challenges. New algorithms and tools need to be developed to cover all steps of data integration. This thesis examines the established data integration proceses and how they can be applied to Linked Data, with focus on data fusion and conflict resolution. Novel algorithms for Linked Data fusion are proposed and the task of supporting trust with provenance information and quality assessment of fused data is addressed. The proposed algorithms are implemented as part of a Linked Data integration framework ODCleanStore.
|
|
Possibilities of Electrochemical Analysis Using a System of Electrodes With Non-Specific Response
Ederer, Jakub ; Nesměrák, Karel (advisor) ; Ludvík, Jiří (referee)
The master thesis present the possibilities of processing of electrochemical data from a group of four electrodes with non-selective response (simple sensor array) for electrochemical analysis with potential application of the results achieved in the construction of the sensor field type "electronic tongue". This simple system was applied to the sample simulating the food product. Electrochemical data were processed through mathematical operations such as Gaussian approximation, deconvolution or using basic mathematical operations.
|
|
Automatic detection of driving lanes geometry based on aerial images and existing spatial data
Růžička, Jakub ; Brůha, Lukáš (advisor) ; Brodský, Lukáš (referee)
The aim of the thesis is to develop a method to identify driving lanes based on aerial images and existing spatial data. The proposed method uses up to date available data in which it identifies road surface marking (RSM). Polygons classified as RSM are further processed to obtain their vector line representation as the first partial result. While processing RSM vectors further, borders of driving lanes are modelled as the second partial result. Furthermore, attempts were done to be able to automatically distinguish between solid and broken lines for a higher amount of information contained in the resulting dataset. Proposed algorithms were tested in 20 case study areas and results are presented further in this thesis. The overall correctness as well as the positional accuracy proves effectivity of the method. However, several shortcomings were identified and are discussed as well as possible solutions for them are suggested. The text is accompanied by more than 70 figures to offer a clear perspective on the topic. The thesis is organised as follows: First, Introduction and Literature review are presented including the problem background, author's motivation, state of the art and contribution of the thesis. Secondly, technical and legal requirements of RSM are presented as well as theoretical concepts and...
|
|
Collision Avoidance For Ateros Robotic System
Ligocki, Adam
This paper describes the details of a collision avoidance algorithm for an ATEROS robotic system. The solution, developed and tested on the Orpheus robotic platform is based on a Velodyne HDL-32E laser scanner. The LiDAR point cloud input data are filtered to remove data redundancy and clustered to separate possible collision objects from the background. Based on prior environment knowledge and the current LiDAR scan, the surrounding occupancy grid map is estimated, and the planned path is validated against possible collision. In the case of a non-zero probability that the robot collides with an obstacle, a new path is proposed by the A* algorithm. Subsequently, the newly estimated waypoints are relaxed, and the mission plan is updated.
|
|
Robot Positioning Based on Sensor Measurements
Čakloš, Ondrej ; Žák, Marek (referee) ; Rozman, Jaroslav (advisor)
The goal of this thesis is to create a program, that will be receiving measurements from robot's sensors and fuse them together. Afterwards use this data fusion of chosen sensors to estimate location of a robot. As a solution for these problems I have used my knowledge of Kalman filters, especially extended one. If messages from sensor measurements are well formulated, Kalman filter can perform fusion of measurements together with estimating the actual position of a robot. Filter can receive measurements from multiple sources and even from duplicities. The estimated states have proven themselves reasonably accurate for successful robot localization in space.
|