National Repository of Grey Literature
Not a member yet
574846 research outputs found
Sort by
Analysis of the Effectiveness of Reels with Impact on Client Acquisition in Investment Advisory Using BI and AI
No full textThis bachelor thesis focuses on the use of reels videos on social media in the field of investment consulting. Using Power BI tools and basic artificial intelligence features, the performance of videos created for the company Kratochvíla & Partners was analyzed. It focuses on reach and viewership to determine how different forms of content influence the behavior of the target audience. The outcome includes an evaluation of the approach and subsequent recommendations
Grate Wood Chips and RDF Boiler
No full textThe thesis focuses on the design of a grate steam boiler with an output of 40 t/h designed to combust a mixture of wood chips and waste (RDF) in a 60/40% ratio. The proposed parameters of superheated steam at the outlet are: temperature 435 °C and pressure 4,2 MPa, while the feedwater temperature is set at 105 °C. The introductory part of the thesis includes stoichiometric calculation of the boiler and determination of its efficiency by indirect method. This is followed by a detailed thermal calculation of the individual passes and heat transfer surfaces of the boiler. The appendix of the thesis contains a dimensional design of the boiler including basic drawings and layout. The final part of the thesis is devoted to the evaluation and control of the overall heat balance of the boiler, determination of pressure losses on the flue gas and medium side and assessment of the risk of low and high temperature corrosion
Information Modelling
No full textThis thesis is divided into two main parts. The first part focuses on mobile scanning using SLAM technology and evaluates the accuracy of the acquired data. The second part describes the application of the requirements of the Czech Construction Data Standard, issued by the Czech Agency for Standardization, to the information model of one floor of the Institute of Geodesy. The results of both parts are appropriately visualized and commented on
Two-stage and bilevel programming
No full textTato diplomová práce se zabývá dvěma typy optimalizačních modelů: stochastickým dvoustupňovým modelem a dvouúrovňovým modelem, a jejich deterministickými, resp. stochastickými variantami. Práce pojednává o již známých vztazích mezi těmito modely a dále je rozšiřuje. Především se v práci ukazuje ekvivalence mezi Karush-Kuhn-Tucker (KKT) jednostupňovými reformulacemi dvouúrovňového a stochastického dvouúrovňového modelu. Tato ekvivalence dovoluje použít algoritmy navržené pro deterministické dvouúrovňové modely i na obecnější stochastické dvouúrovňové modely, a to s žádnými nebo pouze malými úpravami těchto algoritmů. Dále pak lze upravené dvouúrovňové algoritmy testovat při řešení stochastických dvoustupňových úloh. Nejprve je však představen přehled algoritmů pro dvoustupňové a dvouúrovňové modely, ze kterého jsou vybrány čtyři algoritmy k implementaci — lineární a nelineární pro každý typ modelu. Tyto čtyři algoritmy jsou následně podrobněji diskutovány a je představena modifikace dvouúrovňových algoritmů pro použití ve scénářovém stochastickém rámci. Praktická použitelnost dvouúrovňových algoritmů na vybrané dvoustupňové úlohy je posouzena prostřednictvím analýzy složitosti spojené s reformulací dvoustupňové úlohy do tvaru vhodného pro řešení pomocí dvouúrovňových algoritmů. Poté jsou algoritmy testovány na vybrané lineární a nelineární dvoustupňové úloze. Toto testování slouží k odhalení implementačních problémů, které by nemusely být zřejmé z teoretických zjištění. Vzhledem k tomu, že algoritmy navržené přímo pro stochastické dvoustupňové úlohy by měly dosahovat lepších výsledků, testování výkonu sloužilo jako srovnání rozdílů ve výkonnosti. Kromě toho porovnání jednotlivých principů metod řešení společně s porovnáním struktur úloh vedlo k odhalení cest pro další výzkum v oblasti stochastických dvouúrovňových algoritmů
Mobile Application Security Scenario for the BUTCA Education Platform
No full textThis thesis focuses on the design and implementation of two educational scenarios aimed at mobile application security, intended for use within the BUTCA (Brno University of Technology Cyber Arena) training platform. The first scenario targets penetration testing of a deliberately vulnerable Android application, demonstrating common security flaws based on the OWASP Mobile Top 10 methodology. The second scenario addresses incident handling and mobile application security analysis using forensic tools, malware detection, and network traffic analysis. The thesis also includes a theoretical overview of current mobile security threats, methodologies, and testing techniques. The proposed scenarios were designed to support practical, hands-on learning and enhance cybersecurity training. Following implementation, the scenarios were tested and evaluated using analytical tools available in the BUTCA platform. The results demonstrate their value in cybersecurity education, particularly in developing students’ ability to identify, analyze, and respond to mobile application threats effectively
Analysis of GaN dislocations by microscopic techniques
No full textTato práce se zabývá vývojem škálovatelného systému pro analýzu vláknových dislokací v nitridu gallia (GaN). V rámci práce byla navržena a implementována automatizovaná metoda pro detekci a analýzu jednotlivých defektů pomocí zobrazování kontrastu elektronového kanálování (ECCI) ve skenovacím elektronovém mikroskopu. Současně byl vytvořen algoritmus pro detekci míst úniku proudu v datech z vodivostní mikroskopie atomárních sil (CAFM). Byly provedeny korelativní ECCI-CAFM experimenty na průmyslových AlGaN/GaN epivrstvách, pro účely CAFM měření pak byla vyvinuta šetrná metoda pro dekontaminaci povrchu vzorků. Předběžná statistická analýza souboru 12~000 dislokací naznačuje existující souvislost mezi vodivostí defektů a jejich strukturálními znaky viditelnými v ECCI. Výsledky této práce vytvářejí pevný základ pro rozsáhlé korelační studie dislokací v GaN, které v budoucnu mohou přispět ke zlepšení účinnosti a spolehlivosti polovodičových součástek pro vysokovýkonové aplikace
Identifying Activities in Video
No full textThis bachelor's thesis focuses on the recognition of human activities in video recordings using computer vision methods. The aim is to design and implement a system that can automatically detect people, track their movement, and classify their activities based on camera footage without the need for manual intervention. The system leverages modern computer vision techniques, particularly object detection, object tracking, and activity recognition using machine learning. The thesis describes the overall system design as well as its individual functional components. The results show that the proposed system is capable of successfully identifying predefined activities in video footage from a street environment. The final part of the thesis discusses the limitations of the solution and suggests possibilities for further development
Tools and Methodical Approach for Multimodal Waste Transport Modelling
No full textWith the increasing amount of waste produced by today's society, it is essential to focus on making its treatment more efficient, reducing its production, and optimizing waste management overall. Waste transport plays a crucial role in this process. While significant attention is paid to optimizing waste collection, less focus is given to transporting waste over longer distances, especially when using multimodal transport. Current research and models often simplify many aspects, exclude parts of the chain, or lack connection to real infrastructure. This work aims to develop a tool for evaluating multimodal transport (road and rail), addressing the shortcomings of existing models. It allows for the evaluation of not only the entire transport chain but also its individual parts. The proposed techno-economic model of multimodal transport primarily focuses on transport within waste management. However, it is designed as a general tool for evaluating a wide range of transport scenarios. The model integrates the road and rail network based on the real infrastructure of the Czech Republic and includes an extensive database of transport system elements, from which user can modularly assemble a transport system suitable for specific scenarios. The technical evaluation includes determining weight, finding transport route, calculating theoretical fuel/energy consumption, and related exhaust gas emissions, as well as detailed time calculations. The subsequent economic evaluation provides a detailed breakdown of costs and estimates the distribution of income and expenses over the project's duration. The model was verified through several case studies, and a methodology for using the model and integrating it into solved cases was established, which was presented on selected topics
Design of Local Cryotherapy Device
No full textThe topic of this diploma thesis is the design of a device intended for the application of extreme cold to a specific area of a patient’s skin. The device is designed for use in medical environments. The design is based on both technical and user analysis. An innovative feature of the proposal is the solution for transporting the liquid nitrogen container, which can be easily moved using an integrated trolley that is created by detaching the rear part of the device. Thanks to a simple connection system between these parts, transportation is quick and easy. The construction is adapted for outdoor transport of the trolley, including overcoming obstacles such as door thresholds
Automatic image optimization
No full textPhotos from traffic cameras often exhibit negative properties regarding their presentation. This thesis compares and combines methods for their improvement, while maintaining the visual continuity of the image. The main goal is the enhancement of vehicles, suppression of noise and the elimination of artefacts. Additionally, an algorithm for the automatic correction of motion blur of the images is described. Among the main methods used are the brightness correction using specific histograms, equalization of local histograms, filtering using windows in the frequency transform of the image, bilateral filtering, gamma correction and Wiener filtering. To achieve desired results segmentation methods based on edge detection are used and subsequent merging of photos using Poisson image reconstruction