|
|
Dataset augmentation with style transfer methods
Wolny, Michał ; Ligocki, Adam (referee) ; Kratochvíla, Lukáš (advisor)
This bachelor's thesis focuses on the research of dataset augmentation and style transfer methods. From the range of available style transfer algorithms, three very different methods were selected, implemented and then experimentally used for dataset augmentation. The effectiveness of augmentation using these methods was verified by performing a statistical analysis of each newly created dataset compared to the original, unmodified dataset. The results of the analysis provide important information about changes in statistical characteristics such as entropy, mean, median, variance, and standard deviation. This information helped to evaluate the effectiveness and impact of the augmentation methods used on the augmented dataset and provide evidence of their potential.
|
|
| |
|
|
Approximation of functions determining colony activity using neural networks
Nevláčil, Jakub ; Ligocki, Adam (referee) ; Honec, Peter (advisor)
Bees as a primary pollinator are an indispensable contribution to global agriculture and food production. However, their numbers have been constantly declining in recent times, primarily due to climate change, parasites or the effect of pesticide use. Understanding their behavior and reliably determine their activity and health could significantly prevent or slow down their decline. That is why this work deals with the development of a device for the acquisition of useful data from beehives, which could be used to determine the activity and health of the bees. Furthermore, this work deals with analysis of the accumulated data using machine learning methods with an emphasis on determining the activity and health of the bees.
|
|
|
LiDAR-based Indoor Self Localization and Mapping
Minařík, Jakub ; Gábrlík, Petr (referee) ; Ligocki, Adam (advisor)
This thesis introduces the problem of simultaneous localization and mapping (SLAM) with focus on the use of a 3D LiDAR sensor. Firsly is written an introduction to SLAM itself and explained graph-based SLAM and dense map representation. The two most common point cloud alignment algorithms ICP and NDT are described. Then research of existing projects solving this problem is carried out. Described projects are all open-source and most of them support the ROS system. One of the described projects, Optimized SC-F-LOAM is explained in detail. Thesis describes it's odometry FLOAM and connection between it and graph optimization with loop closure. For loop closure is used descriptor ScanContext. Then it is presented design for implementing choosen project on offline and online datas from indoor. In last chapters is described proces of implementing and tuning project and at the end results of using said project in indoors are presented.
|
|
|
Path planning for multi-agent robotic systems
Macák, Libor ; Ligocki, Adam (referee) ; Lázna, Tomáš (advisor)
This work concentrates on developing a Robot Operating System 2 package that enables multiagent path planning and puts main emphasis on non-collision traffic between more agents at the same time. Next it explains some basic concepts about Robot Operating System 2. This work also approaches multiagent path planning problem with listing some basic theoretical concepts and it specifies some used algorithms.
|
|
|
Implementing the checkers game using a robotic manipulator
Lichosyt, David ; Ligocki, Adam (referee) ; Lázna, Tomáš (advisor)
The thesis starts with a field research about existing robotic systems with implemented checkers or similar desk games. In its next parts are described posibilities of controling robotic manipulator Fanuc LR Mate 200iD 4S using controller R-30iB Mate from platform ROS located on external hardware. In few next pages of the thesis there are mentioned designed and used physical models. Following chapters describe real time game situation detection, game logic implementation and human input with safety in mind. At the end of the thesis there are mentioned rule checking algorithms, system operation instructions and possible future upgrades of the system.
|
|
|
Machine Learning-Based Multimodal Data Processing and Mapping in Robotics
Ligocki, Adam ; Duchoň,, František (referee) ; Saska,, Martin (referee) ; Žalud, Luděk (advisor)
Disertace se zabývá aplikaci neuronových sítí pro detekci objektů na multimodální data v robotice. Celkem cílí na tři oblasti: tvorbu datasetu, zpracování multimodálních dat a trénování neuronových sítí. Nejdůležitější části práce je návrh metody pro tvorbu rozsáhlých anotovaných datasetů bez časové náročného lidského zásahu. Metoda používá neuronové sítě trénované na RGB obrázcích. Užitím dat z několika snímačů pro vytvoření modelu okolí a mapuje anotace z RGB obrázků na jinou datovou doménu jako jsou termální obrázky, či mračna bodů. Pomoci této metody autor vytvořil dataset několika set tisíc anotovaných obrázků a použil je pro trénink neuronové sítě, která následně překonala modely trénované na menších, lidmi anotovaných datasetech. Dále se autor v práci zabývá robustností detekce objektů v několika datových doménách za různých povětrnostních podmínek. Práce také popisuje kompletní řetězec zpracování multimodálních dat, které autor vytvořil během svého doktorského studia. To Zahrnuje vývoj unikátního senzorického zařízení, které je vybavené řadou snímačů běžně užívaných v robotice. Dále autor popisuje proces tvorby rozsáhlého, veřejně dostupného datasetu Brno Urban Dataset. Na závěr autor popisuje software, který vznikl během jeho studia a jak je tento software užit při zpracování dat v rámci jeho práce (Atlas Fusion a Robotic Template Library).
|
|
|
Evaluation Of The Neural Network Object Detection In Multi-Modal Images
Ligocki, Adam
This paper studies the information gain of various data domains that are commonly usedin the modern Advanced Driving Assistant Systems (ADAS) to develop robust systems that wouldincrease traffic safety. We could see a fast growth of many Deep Convolutional Neural Networks(DCNN) based solutions during the last several years. These methods are state-of-the-art in objectdetection and semantic scene segmentation. We created a small annotated dataset of synchronizedRGB, grayscale, thermal, and depth map images and used the modern DCNN framework tool toevaluate the object detection robustness of different data domains and their information gain processunderstanding the surrounding environment of the semi-autonomous driving agent.
|
|
|
Transfer Learning For Deep Convolutional Neural Network From Rgb To Ir Domain
Ligocki, Adam ; Jelínek, Aleš
In this paper, we are presenting a proof of concept of our system for training of the YOLOv3 neural network for object detection of vehicles in thermal camera images. Our approach is unique in the way we are using a dataset containing a large number of synchronized range measurements as well as RGB and thermal images. We are using the existing YOLO toolkit to detect objects on the RGB images, we estimate detection distance by the LiDAR and later we reproject these detections into the IR image. In this way, we have created a large dataset of annotated thermal images that helped us to significantly improve the performance of the neural network at the IR domain.
|
|
|
Peristaltic Pump Controll Unit
Charvát, Jakub ; Kopečný, Lukáš (referee) ; Ligocki, Adam (advisor)
The work deals with the problematic of peristaltic pumps and it collects theoretical background information of this topic. The primary goal of the work is the implementation of peristaltic pump and the implementation of protection against equipment damage by overpressure. The overpressure evaluation system in this thesiss on the principle of current measurement of the motor pump. In this work, this premise is successfully tested and applied to the device. Another goal of this work is to create a desktop application that will monitor and control the pump. he data between the application and the device were successfully transfered by the MQTT communication protocol. The communication between the computer and the pump has been tested successfully.
|
guest ::
