Syftet med examensarbetet är att undersöka markstödens placeringar och fördelningar i plan och om höjden har en inverkan på noggrannheten i ortomosaik. En jämförelse genomförs mellan totalstation och UAV-koordinater. Först utfördes rekognosering i fält för att sedan placera ut åtta stycken markstödsplattor varpå ett av markstöden blev placerad på en högre höjd. En ytterligare variant av markstödstyp placerades ut i form av ett grönt kryss med sprayflaska. Markstöden och kontrollpunkterna mättes in med totalstation efter fristationsuppställning där tre bakåtobjekt används för att erhålla noggranna plana koordinater. En fixpunkt används för att erhålla stabila höjder.  I denna studie användes en UAV i form av DJI Phantom 4 Pro vid flygfotograferingen och en totalstation (Trimble S7) för inmätningarna. Inmätningen utfördes med totalstation mot kontrollpunkterna med samtliga markstöd som tillämpades sedan som facit. Slutligen utfördes UAV-flygning med en flygrutt på 60 m höjd och 90 graders vinkel för UAVkameran. Bildbearbetningen utfördes i programvaran Agisoft Metashape där fem stycken modeller genererades i form av ortomosaik och DEM. Sedan med hjälp av ortomosaiken exporterades koordinaterna till programvaran Microsoft Excel för vidare bearbetning och kontroll av avvikelser mellan UAV- och totalstationens koordinater. Resultaten från ortomosaikmodellerna visar att RMS i plan och höjd påverkas beroende på antalet markstöd. Modellen utan markstöd resulterar i en stor koordinatdifferens med RMS i plan på cirka 2 m. Vid användning av fyra markstöd ses en markant förbättring i noggrannhet för RMS i plan på cirka 0,008 m och i höjd på 0,045 m. För modellen med fem markstöd påvisas inga märkbara skillnader för RMS i plan, men för RMS i höjd erhålls betydligt lägre mätosäkerhet. Sista modellen med åtta markstöd resulterar i lägst mätosäkerhet i plan och höjd jämfört mot de andra modellerna, men skillnaden var oansenlig mellan fem och åtta markstöd. Utifrån frågeställningen och framtaget resultat kan det fastslås att ortomosaiksmodellen med fem markstöd är lämpligast för vidare användning vid flygfotografering. Markstödsplatta rekommenderas för noggrannare höjdkoordinater jämfört med märkspray. Om endast plana koordinater ska bearbetas kan antingen märkspray eller markstödsplatta utnyttjas för att erhålla likvärdigt resultat. / The objective of the bachelor thesis is to examine if ground control point (GCP) placing and allocation would impact the accuracy of an orthomosaic. A comparison between total station and UAV-coordinates is carried out. Reconnaissance was carried out in the field to deploy the GCP around the study area with one of them deployed on top of a pile. A different variant of GCP is deployed in the form of green-coloured cross that was applied with the help of a spray bottle. Both GCP and control points were measured with total station after the free station set-up where the three backsights and a fixed point were used to obtain a stable height. During the studies, a DJI Phantom 4 Pro UAV was used for aerial photography and a total station (Trimble S7) for the measurements. Before the flight, measurements were calculated with the total station to the control points and all GCP, which were then used as a record. GNSS-drone flying was performed with a flight route at a height of 60 meters and a 90 degrees angle for the UAV camera. Image processing was performed in software Agisoft Metashape where a total five orthomosaic models were generated. Then with the help of orthomosaic, the coordinates were transferred to the software Microsoft Excel for further processing and control of deviation between the coordinates of the UAV and the total station. Results from the orthomosaic models show that RMS in both plane and height do impact accuracy depending on the numbers of GCP used. A model without the use of GCP shows a huge coordinate difference for RMS in plane at approximately two meters. The use of four GCP models show a significant improvement in both RMS in plane at approximately 0,008 meters and the height at 0,046 meters. Five GCP models show a slight difference for RMS in plane but is considerable when the height was at 0,008 meters. Lastly, eight GCP models show some small significant improvements in both RMS plane and height.  Based on the questions and results, it can be stated that the orthomosaic models with five GCP are recommended for further use of drone surveying. The GCP wooden plate is to be preferred for a more accurate height compared to the green cross spray method. But if only planar coordinates are considered, both GCP wooden plate and green cross spray bottle can be used with equal accuracy.

The popularity of using UAV in image-taking for the production of 3D models and orthophotos has increased over time. Karlskoga Municipality has recently acquired an UAV to produce their own 3D models and orthophotos. This project paper aims to study the geospatial accuracy of the orthophotos and DEM files produced using the images taken with their UAV. The flight takes only a few minutes but a considerable time is spent in the production processes. Difficulty is experienced in determining the right center point for most GCPs. Produced orthophotos in the software Photoscan have a resolution from 1.7 to 2.4 centimeters while DEM files have a resolution from 3.4 to 4.8 centimeters. Four orthophotos and four DEM files are produced where GCPs are used and not used and at two different flight heights, 76 and 105 meters. The spatial data of the ten GCPs are identified on the orthophotos and DEM files in ArcMap and compared with GNSS NRTK measurements and Lantmäteriet's data. A visual control in terms of completeness of data, alignment, residual tilt and scale is also done. Standard deviations in plane for orthophotos there GCPs are not used are greater than 2 meters, while there GCPs are used are around 0.7 meters. Standard deviations for DEM files are observed at 0.8 meters.

Cloud computing offers higher reliability and lower up-front IT costs than traditional computing environments and is a great way to dynamically scale both resources and capabilities. For further efficiency and consistency, cloud computing tasks can also be automated with tools such as Ansible. In this thesis, we will analyze and compare the abilities of Ansible with the three leading cloud platforms, i.e. Azure, Amazon Web Services (AWS), and Google Cloud Platform (GCP). To evaluate this, we cover three different areas of automation with cloud platforms. These areas are performance, user complexity, and missing network functionalities. The performance was evaluated through experiments that revealed a big gap between the platforms, where AWS was the clear winner in all scenarios. Microsoft Azure was slightly faster than GCP when a low number of virtual machines were created but GCP scaled better than Azure when more virtual machines were created. The user complexity was evaluated on the setup process and the creation of playbooks where Azure was the clear winner in both areas. AWS and GCP had similar setup processes, but AWS takes second place through its superior documentation for the creation of playbooks. All three platforms had missing network functionalities, but the majority of the missing functionalities were not relevant as they were not related to deployment which is the main usage of an automation tool like Ansible. However, some deployment functions were missing, for example, the firewall function for AWS was missing in Ansible.

In order to fully benefit from cloud computing, services are designed following the “multi-tenant” architectural model which is aimed at maximizing resource sharing among users. However, multi-tenancy introduces challenges of security, performance isolation, scaling and customization. RStudio server is an open source Integrated Development Environment (IDE) accessible over a web browser for R programming language. The purpose of this thesis is to develop an open source multi-user distributed system on Hopsworks, a data intensive and AI platform, following the multi-tenant model, that provides RStudio as Software as a Service (SaaS). Our goal is to promote collaboration among users when using RStudio and the learning and teaching of R by enabling users easily have access to same computational environments and resources while eliminating installation and maintenance tasks. Hopsworks introduces project-based multi-tenancy where users within a project share project resources (e.g datasets, programs and services) for collaboration which introduces one more challenge of sharing project resources in RStudio server instances. To achieve our purpose and goal we therefore needed to solve the following problems: performance isolation, security, project resources sharing, scaling and customization. Our proposed model is on demand single user RStudio server instances per project. Our system is built around Docker and Kubernetes to solve the problems of performance isolation, security and scaling. We introduce HopsFS-mount, that allows securely mounting HopsFS via FUSE to solve the project resources (datasets and programs) sharing problem. We integrate our system with Apache Spark which can scale and handle Big Data processing workloads. Also we provide a UI where users can provide custom configuration and have full control of their own RStudio server instances. Our system was tested on a GCP cluster with four worker nodes each with 30GB of RAM allocated to them. The tests on this cluster showed that 44 RStudio servers, each with 2GB of RAM, can be run concurrently. Our system can scale out to potentially support hundreds of concurrently running RStudio servers by adding more resources (CPUs and RAM) to the cluster or system. / För att dra full nytta av molntjänster är vissa applikationer designade för multitenans som syftar till att maximera resursdelning mellan användare. Dock introducerar multitenans utmaningar i hänsyn till resursdelning, säkerhet, prestandaisolering, anpassning och skalning. RStudio-server är en öppen källkod Integrerad utvecklingsmiljö (IDE) tillgänglig över en webbläsare för programmeringsspråket R. Syftet med denna avhandling är att utveckla ett distribuerat system med öppen källkod för flera användare på Hopsworks, en data krävande AI-plattform, efter multitenans-modellen, som tillhandahåller RStudio som Software as a Service (SaaS). Vårt mål är att främja samarbete mellan användare vid användning av RStudio, inlärning och undervisning av R genom att göra det enkelt för användare att ha tillgång till samma beräknings miljöer och resurser samtidigt som installation och underhållsarbete elimineras. Hopsworks introducerar projektbaserad multitenans där användare inom ett projekt delar projektresurser (t.ex. datamängder, program och tjänster) för samarbete som introducerar ytterligare en utmaning att dela projektresurser i RStudio server instanser. För att uppnå vårt syfte och mål behövde vi därför lösa följande problem: prestandaisolering, säkerhet, projekt resursdelning, skalning och anpassning. Vår föreslagna modell är på bergäran en-användares RStudio-serverinstanser per projekt. Vårt system är byggt kring Docker och Kubernetes för att lösa problemen med prestanda isolering, säkerhet och skalning. Vi introducerar HopsFS-mount, som gör det möjligt att säkert montera HopsFS via FUSE för att lösa resurs (datamängder och program) delning problemet. Vi integrerar vårt system med Apache Spark som kan skala och hantera Big Data bearbetning belastningar. Vi tillhandahåller också ett användargränssnitt där användare kan tillhandahålla anpassad konfiguration och ha full kontroll över sina egna RStudio-serverinstanser. Vårt system testades på ett GCP-kluster med fyra arbets noder, varje node hade 30 GB RAM. Testerna på detta kluster visade att 44 RStudio-servrar, var och en med 2 GB RAM, kan köras samtidigt. Vårt system kan även skala ut för att potentiellt stödja hundratals RStudio-servrar som samtidigt körs genom att lägga till fler resurser (CPU:er och RAM) i klustret eller systemet.

The continued cost reduction for sequencing genomics data is causing an exponentialgrowth in the amount of data available. Moving both storage and calculation of thisdata to the cloud has been a common trend, but the way to do it is not alwaysobvious. This report compares three different alternatives for doing ad-hoc queries ina cloud based setting: two solutions using data lakes and one solution using arelational database hosted in the cloud. The data lake solutions proved to be easy toset up and fully functional for querying genomics data. The relational database wasmore complicated to set up, but the queries were more time efficient and more costefficient when performing more than 1200 queries per month on at least 100GB ofdata. To make the cloud computing possible for genomics data it had to betransformed into a file format supported by the cloud providers. For this purpose theParquet file format was chosen, tested, and proven to work well

Conducting single frame orthorectification on satellite images to create an ortho-image requires four basic components, namely an image, a geometric sensor model, elevation data (for example a digital elevation model (DEM)) and ground control points (GCPs). For this study, orthorectification was executed numerous times (in three stages) and each time components were altered to test the geometric accuracy of the resulting ortho-image. Most notably, the distribution and number of ground control points, the quality of the elevation source and the geometric sensor model or lack thereof were altered. Results were analysed through triangulating and comparing the geolocation accuracy of the ortho-images. The application of these different methods to perform orthorectification encompass the aim of this paper, which was to investigate and compare the positional accuracies of ortho-images under various orthorectification scenarios and provide improved geometric accuracies of VHR satellite imagery when diverse ground control and elevation data sources are available. By investigating the influence that the distribution and number of GCPs and the quality of DEMs have on the positional accuracy of an ortho-image, it became clear that a reasonable increase in the number of uniformly distributed GCPs combined with progressively accurate DEMs will ultimately improve the quality of the orthorectified product. The results also showed that when more GCPs were applied, the smaller the difference in accuracy was between the different DEMs utilised. It was interesting to note that when it is suitable to manually collect well-distributed GCPs using a GPS handheld device over the study area then a very accurate result can be expected. Nonetheless, it is also important to note that if it is not possible/practical to achieve the latter, satellite based GCP collection do provide a very good alternative. It was also determined that utilising GCPs which were extracted from vector road layers to only cover specific areas in the image scene produced less favourable results. Several contributions towards improved orthorectification procedures were made in this study. These include the development of an automatic GCP extraction script (A-GCP-ES), written in the Python scripting language with the purpose to ease the process of manually placing GCPs on an input image when repeatedly performing orthorectification. / Thesis (PhD)--University of Pretoria, 2016. / Geography, Geoinformatics and Meteorology / PhD / Unrestricted

Context: In the field of computer science and communication systems, cloud computing plays animportant role in Information and Technology industry, it allows users to start from small and increase resources when there is a demand. AWS (Amazon Web Services) and GCP (Google cloud Platform) are two different cloud platform providers. Many organizations are still relying onstructured databases like MySQL etc. Structured databases cannot handle huge requests and data efficiently when number of requests and data increase. To overcome this problem, the organizations shift to NoSQL unstructured databases like Apache cassandra, Mongo DB etc. Conclusions: From the literature review, I have gained knowledge regarding the cloud computing, problems existed in cloud, which leads to setup this research in evaluating the performance of cassandra on AWS and GCP. The conclusion from the experiment is that as the thread count increases throughput and latency has increased gradually till thread count 600 in both the clouds. By comparing both the clouds throughput values, AWS scales up compare to GCP. GCP scales up, when compared to AWS in terms of latency.  Keywords: Apache Cassandra, AWS, Google Cloud Platform, Cassandra Stress, Throughput, Latency

UAS has been become a very popular tool in surveying and evaluation of the systems measurement uncertainties are necessary. The most common method for georeferencing UAS data is to use ground control points (GCP) in order to use them in block adjustment. In recent years’ new techniques for direct georeferencing with UAS have been presented, which in theory means that the position of the UAS can be determined accurately enough and therefore GCP’s can be excluded. This study evaluates  uncertainties of the UAS Freya from SmartPlanes that don’t need GCP’s for georeferencing. The technique applied in the evaluation is based on Post Processed Kinematic (PPK) for coordinate determination of the UAS, which means that the collected GNSS data can be post processed using a reference station.   The test area was a 280 x 320 m block in the north end of Gävle airport, Sweden. Each flight is conducted in two orthogonal blocks and evaluated in three different ways against the 16 GCP. The altitude was about 90 m for all flights. The uncertainty of the PPK-technique is tested and evaluated with three different methods to ensure both accuracy and potential use. In total five flights were assessed and evaluated with Agisoft PhotoScan against 16 GCP spread over the area. The position of each GCP’s was determined with four independent network RTK measurements.  The results show that the georeferencing with the PPK-technique and block adjustment has potential to meet the uncertainties in level with indirect georeferencing using GCP. The results show very similar planimetric uncertainties, around 0,020 m in RMS, for all evaluations with the PPK-technique. The results of the uncertainty in height is more scattered where the two lowest results in a RMS under 0,015 m and the highest over 0,100 m for the difference against the 16 GCP.  It is possible to achieve low uncertainties with the method without the use of GCP. For areas where establishment of GCP is not possible, using UAS equipped with PPKtechnology provides a very suitable alternative to use. The results show relatively large differences between the evaluations and in order to determine the exact cause of them, further studies are required. / Den starka teknikutvecklingen för UAS resulterar i flera nya produkter på marknaden och för att utvärdera deras mätosäkerheter krävs det kontroller av systemen. Den vanligaste metoden vid georeferering med UAS är att använda koordinatbestämda flygsignaler på marken. På senare år har metoder för direkt georeferering presenterats, vilket i teorin innebär att positionen för UAS kan bestämmas så pass noggrant att flygsignaler kan uteslutas. I denna studie utvärderas mätosäkerheter för Freya, ett UAS från SmartPlanes som med hjälp av bra positionering och blockutjämning ska kunna användas för georeferering utan flygsignalering. Systemet från Smartplanes bygger på Post Processed Kinematic (PPK) för koordinatbestämning, vilket innebär att insamlat GNSS data kan efterberäknas med korrektioner från en referensstation. Mätosäkerheten för PPK-tekniken testas och utvärderas med 3 olika metoder för att se både på mätosäkerhet samt möjlig användning. Totalt fem flygningar har utförts på ett testområde som var cirka 280x320 m och beläget i den norra delen av Gävle flygplats. Flyghöjden var kring 90 m för alla flygningar som vidare har bearbetats och utvärderats i programvaran PhotoScan från Agisoft. Kontrollen av mätosäkerheten har gjorts mot 16 spridda kontrollpunkter på marken som har positionsbestämts med fyra oberoende nätverks-RTK mätningar vardera. Varje flygning är utförd i två ortogonala block och utvärderades med fyra olika konfigurationer mot de 16 kontrollpunkterna.  Resultaten visar att georeferering med hjälp av blockutjämning och PPK-tekniken har potential för att uppnå mätosäkerheter i nivå med indirekt georeferering med hjälp av stödpunkter på marken. I plan visar resultaten på väldigt jämna mätosäkerheter, kring 0,020 m i RMS, för alla utvärderingar med PPK-tekniken. Resultaten i höjd är mer spridda där de lägsta visar mätosäkerheter under 0,015 m RMS och de högsta över 0,100 m i RMS för avvikelsen mot kontrollpunkterna. Det är fullt möjligt att uppnå låga mätosäkerheter med metoden utan användning av stödpunkter. För användningsområden där stödpunkter inte kan etableras är detta UAS med PPK-tekniken ett mycket lämpligt alternativ att använda. Resultaten visar på relativt stora skillnader mellan de olika testade metoderna och för att avgöra den exakta orsaken till dem skulle vidare studier behövas.

Generation of Digital Surface Models (DSMs) from stereo satellite (spaceborne) images is classically performed by Ground Control Points (GCPs) which require site visits and precise measurement equipment. However, collection of GCPs is not always possible and such requirement limits the usage of spaceborne imagery. This study aims at developing a fast, fully automatic, GCP-free workflow for DSM generation. The problems caused by GCP-free workflow are overcome using freely-available, low resolution static DSMs (LR-DSM). LR-DSM is registered to the reference satellite image and the registered LR-DSM is used for i) correspondence generation and ii) initial estimate generation for 3-D reconstruction. Novel methods are developed for bias removal for LR-DSM registration and bias equalization for projection functions of satellite imaging. The LR-DSM registration is also shown to be useful for computing the parameters of simple, piecewise empirical projective models. Recent computer vision approaches on stereo correspondence generation and dense depth estimation are tested and adopted for spaceborne DSM generation. The study also presents a complete, fully automatic scheme for GCPfree DSM generation and demonstrates that GCP-free DSM generation is possible and can be performed in much faster time on computers. The resulting DSM can be used in various remote sensing applications including building extraction, disaster monitoring and change detection.

Darbo tikslas: Atskleisti organizacijos kultūros ir visuotinės kokybės vadybos principų dermę klinikinių tyrimų organizacijoje. Darbo uždaviniai: 1. Ištirti klinikinių tyrimų organizacijos kultūrą ir visuotinės kokybės vadybos principų dermę. 2. Vadovaujantis Denison organizacijos kultūros modeliu, išanalizuoti klinikinių tyrimų organizacijos kultūros ypatumus. 3. Pateikti rekomendacijas dėl klinikinių tyrimų organizacijos kultūros tobulinimo. / Aim of the research – to reveal the coherence of organization culture and total quality management principles in clinical research organization. Targets: 1. To investigate clinical research organization culture and the coherence of total quality management. 2. To introduce clinical research organization. 3. According to the model of Denison to analyze organizational culture characteristics of clinical research organization. 4. To provide recommendations for the improvement of clinical research organization culture.