|
|
Reconstruction of AXMAT Testing Station
Gergel, Matej ; Dobeš, Petr (referee) ; Dvořáček, Jiří (advisor)
The master´s thesis deals with reconstruction of RCF test rig AXMAT. Main goal is design new hydraulic load system and his realization. The first part of thesis describes similar experimental machines and their pros and cons. This analysis is base for complete reconstruciton of AXMAT. New hydraulic load system allows dynamic programmable load cycles. Control system with close loop and feedback was created in software Matlab – Simulink. Main frame was reconstructed too. Control and action parts were asseble to station. Output of thesis is function sample.
|
|
|
OCR Trained with Unanotated Data
Buchal, Petr ; Dobeš, Petr (referee) ; Hradiš, Michal (advisor)
The creation of a high-quality optical character recognition system (OCR) requires a large amount of labeled data. Obtaining, or in other words creating, such a quantity of labeled data is a costly process. This thesis focuses on several methods which efficiently use unlabeled data for the training of an OCR neural network. The proposed methods fall into the category of self-training algorithms. The general approach of all proposed methods can be summarized as follows. Firstly, the seed model is trained on a limited amount of labeled data. Then, the seed model in combination with the language model is used for producing pseudo-labels for unlabeled data. Machine-labeled data are then combined with the training data used for the creation of the seed model and they are used again for the creation of the target model. The successfulness of individual methods is measured on the handwritten ICFHR 2014 Bentham dataset. Experiments were conducted on two datasets which represented different degrees of labeled data availability. The best model trained on the smaller dataset achieved 3.70 CER [%], which is a relative improvement of 42 % in comparison with the seed model, and the best model trained on the bigger dataset achieved 1.90 CER [%], which is a relative improvement of 26 % in comparison with the seed model. This thesis shows that the proposed methods can be efficiently used to improve the OCR error rate by means of unlabeled data.
|
|
|
Monitor of Exploitation in Ultramicro Aviation
Kudera, Matěj ; Dobeš, Petr (referee) ; Zemčík, Pavel (advisor)
The aim of this bachelor's thesis was to create a system that will record the operation of an ultralight aircraft in which it will be installed. The thesis analyzes the principles important for the design of this system and creates a detailed proposal, which is based on the acquired knowledge of individual principles. The created system consists of two parts. An embedded device that is built into the ultralight aircraft and records whether it is in motion and a mobile application that allows easy access to records made by the embedded device. The mobile application was made for the Android operating system using the Android Studio development environment and the embedded device was implemented using a Raspberry Pi minicomputer, to which the necessary peripherals were connected. Finally, the functionality of the system was tested using the created tests and test flights of ultralight aircraft with embedded device were performed.
|
|
|
Limiting and Degradation Processes of Magnetorheological Suspension Dampers
Roupec, Jakub ; Bauer, František (referee) ; Ondrůšek, Čestmír (referee) ; Dobeš, Petr (referee) ; Mazůrek, Ivan (advisor)
This work deals with the research of degradation of magnetorheological (MR) fluid during long-term loading under conditions that correspond to the real load in the linear MR devices such as dampers. The behaviour of MR fluid is described by a Bingham model of viscoplastic fluid. The parameters of this model are the yield stress and viscosity. Their values are determined by the shear rate from 1 to 2,8104 s-1. Results of durability tests show a significant decrease in viscosity of MR fluid, a gradual increase in yield stress in the off-state and a significant decrease in MR effect. The paper also described the effect of temperature on these parameters and there is also designed a method of non-assembly diagnostic of the MR fluid state in the damper using the "rate of pulsation."
|
|
|
Study of Elastohydrodynamic Film Formation in Hypoid Gears
Omasta, Milan ; Dobeš, Petr (referee) ; Čermák,, Jan (referee) ; Dzimko, Marián (referee) ; Křupka, Ivan (advisor)
This PhD thesis deals with elastohydrodynamic lubrication (EHL) under the conditions that occur between mated hypoid gear teeth. The aim is to describe experimentally the behaviour of lubricating film and mechanisms of its formation. The focus is mainly placed on clarifying the influence of direction and magnitude of sliding velocity in circular and elliptical contacts. Generally, it is still assumed that the direction of sliding velocity has no effect on the resulting lubricant film. In this study optical interferometry is used to determine distribution of lubricant film thickness. This work includes new and original results which clarify the studied problem. It was found that the direction of sliding velocity affects shape of the film at high speeds. This relates to the effect on heat flow through the contact. The results have an impact on the development of general EHL theory and provide knowledge applicable in film thickness prediction in the design of real tribological nodes.
|
|
|
Reconstruction of 3D Information about Vehicles Passing in front of a Surveillance Camera
Dobeš, Petr ; Sochor, Jakub (referee) ; Herout, Adam (advisor)
This master's thesis focuses on 3D reconstruction of vehicles passing in front of a traffic surveillance camera. Calibration process of surveillance camera is first introduced and the relation of automatic calibration with 3D information about observed traffic is described. Furthermore, Structure from Motion, SLAM, and optical flow algorithms are presented. A set of experiments with feature matching and the Structure from Motion algorithm is carried out to examine results on images of passing vehicles. Afterwards, the Structure from Motion pipeline is modified. Instead of using SIFT features, DeepMatching algorithm is utilized to obtain quasi-dense point correspondences for the subsequent reconstruction phase. Afterwards, reconstructed models are refined by applying additional constraints specific to the vehicle reconstruction task. The resultant models are then evaluated. Lastly, observations and acquired information about the process of vehicle reconstruction are utilized to form proposals for prospective design of an entirely custom pipeline that would be specialized for 3D reconstruction of passing vehicles.
|
|
| |
|
|
Deep Learning for Object Detection
Pitoňák, Radoslav ; Dobeš, Petr (referee) ; Teuer, Lukáš (advisor)
This thesis analyzes different object detection methods which are based on deep neural networks. In the beginning, the convolutional neural networks are described and commonly used object detection methods are compared. In the following parts, the proposal and implementation of the object detection model trained on the specific dataset are described. In conclusion, the achieved results of this model are discussed and compared with the results of other methods.
|
|
|
Information System and User Interface for Career Fair
Pavlík, Tadeáš ; Bartl, Vojtěch (referee) ; Dobeš, Petr (advisor)
This work focuses on the design and implementation of an information system and user interface for day of career opportunities at Olomouc grammar school. The aim of this thesis is to design an information system and user interface through which the presenters can apply for the event, the organizers can manage the event and interact with the visitors of the event. This thesis focuses on implementing an clear and easy - to - use user interface. The information system is implemented using Laravel framework, PHP language and Bootstrap 3 library. The thesis presents the design and implementation of database structure, use cases, information system and user interface. Testing was carried out using participants from previous years. The resultant information system is ready for deployment. The contribution of this work is simplifying the work of the organizers, providing visitors with information and getting information easily on mobile devices.
|
|
|
Deep Neural Networks for Reinforcement Learning
Košák, Václav ; Dobeš, Petr (referee) ; Hradiš, Michal (advisor)
The paper describes a training environment for training a character how to walk. The environment is implemented in Al Gym by using the PyBullet physical model. Tasks from the environment are solved by using reinforcement learning by the ActorCritic algorithm. Each of the tasks is focused on the fundamental movements of the character. The paper show, which reward functions are used by the algorithm to solve given tasks.
|
guest ::
RSS feed.