Offloading SLAM to the edge/cloud is now becoming an attractive option to greatly decrease device energy usage. The new SLAM solution involves compressing image data on the device before transmission, allowing a further decrease in the network bandwidth when performing SLAM at the edge/cloud. However, lossy compression affects the quality of images, making image features harder to detect and track during visual SLAM operation, impacting localization accuracy. Current visual SLAM implementations assume that images are non-compressed since SLAM is traditionally executed onboard the device to which a camera is directly connected. This thesis work explores the impact of image compression on the localization accuracy of ORBSLAM3, a representative visual SLAM system, and in what way the ORBSLAM3’s modules for feature detection and matching are affected. Methods are proposed that adapt the image bitrates based on the number of features detected and enhance the image brightness for low-light conditions, plus optimizing the internal parameters in SLAM, to improve the robustness of the overall system to image compression. The experiment results show the detailed influence of the impact brought by compression on ORB-SLAM3 and prove the effectiveness of our methods. Also, integrating these methods yields synergistic improvements. While this thesis work primarily addresses the SLAM system’s front-end, future work can target back-end modifications. / Att avlasta SLAM till kanten/molnet blir nu ett attraktivt alternativ för att markant minska enheters energiförbrukning. Den nya SLAM-lösningen innebär att bilddata komprimeras på enheten innan den överförs, vilket möjliggör ytterligare minskning av nätverksbandbredden vid genomförandet av SLAM vid kanten/molnet. Men förlustkomprimering påverkar bildernas kvalitet och gör det svårare att upptäcka och följa bildfunktioner under visuell SLAM-drift, vilket påverkar lokaliseringsnoggrannheten. Nuvarande implementationer av visuell SLAM förutsätter att bilderna inte är komprimerade eftersom SLAM traditionellt utförs ombord på den enhet till vilken en kamera är direkt ansluten. Denna avhandling utforskar effekten av bildkomprimering på lokaliseringsnoggrannheten hos ORB-SLAM3, en representativ visuell SLAM-system, och på vilket sätt ORB-SLAM3: s moduler för funktionssökning och matchning påverkas. Metoder föreslås som anpassar bildbitarna baserat på antalet detekterade funktioner och förbättrar bildens ljusstyrka för svagt ljusförhållanden, samt optimerar de interna parametrarna i SLAM för att öka hela systemets robusthet mot bildkomprimering. Experimentresultaten visar den detaljerade påverkan som kompression har på ORB-SLAM3 och bevisar effektiviteten hos våra metoder. Dessutom ger integration av dessa metoder synergi och förbättringar. Även om denna avhandling primärt fokuserar på SLAM-systemets framsida, kan framtida arbete rikta sig mot bakkantsmodifieringar.

Visual(-Inertial) Odometry (VIO) and Simultaneous Localization and Mapping (SLAM) are hot topics in Computer Vision today. These technologies have various applications, including robotics, autonomous driving, and virtual reality. They may also be valuable in studying human behavior and navigation through head-mounted visual systems. A complication to SLAM and VIO systems could potentially be visual degeneration such as motion blur. This thesis attempts to evaluate the robustness to motion blur of two open-source state-of-the-art VIO and SLAM systems, namely Delayed Marginalization Visual-Inertial Odometry (DM-VIO) and ORB-SLAM3. There are no real-world benchmark datasets with varying amounts of motion blur today. Instead, a semi-synthetic dataset was created with a dynamic trajectory-based motion blurring technique on an existing dataset, TUM VI. The systems were evaluated in two sensor configurations, Monocular and Monocular-Inertial. The systems are evaluated using the Root Mean Square (RMS) of the Absolute Trajectory Error (ATE).  Based on the findings, the visual input highly influences DM-VIO, and performance decreases substantially as motion blur increases, regardless of the sensor configuration. In the Monocular setup, the performance decline significantly going from centimeter precision to decimeter. The performance is slightly improved using the Monocular-Inertial configuration. ORB-SLAM3 is unaffected by motion blur performing on centimeter precision, and there is no significant difference between the sensor configurations. Nevertheless, a stochastic behavior can be noted in ORB-SLAM3 that can cause some sequences to deviate from this. In total, ORB-SLAM3 outperforms DM-VIO on the all sequences in the semi-synthetic datasets created for this thesis. The code used in this thesis is available at GitHub https://github.com/cmannila along with forked repositories of DM-VIO and ORB-SLAM3 / Visual(-Inertial) Odometry (VIO) och Simultaneous Localization and Mapping (SLAM) är av stort intresse inom datorseende (Computer Vision). Dessa system har en variation av tillämpningar såsom robotik, själv-körande bilar och VR (Virtual Reality). En ytterligare potentiell tillämpning är att integrera SLAM/VIO i huvudmonterade system, såsom glasögon, för att kunna studera beteenden och navigering hos bäraren. En komplikation till SLAM och VIO skulle kunna vara en visuell degration i det visuella systemet såsom rörelseoskärpa. Detta examensarbete försöker utvärdera robustheten mot rörelseoskärpa i två tillgängliga state-of-the-art system, DM-VIO (Delayed Marginalization Visual-Inertial Odometry) och ORB-SLAM3. Idag finns det inga tillgängliga dataset som innehåller specifikt varierande mängder rörelseoskärpa. Således, skapades ett semisyntetiskt dataset baserat på ett redan existerande, vid namn TUM VI. Detta gjordes med en dynamisk rendering av rörelseoskärpa enligt en känd rörelsebana erhållen från datasetet. Med denna teknik kunde olika mängder exponeringstid simuleras.  DM-VIO och ORB-SLAM3 utvärderades med två sensor konfigurationer, Monocular (en kamera) och Monokulär-Inertial (en kamera med Inertial Measurement Unit). Det objektiva mått som användes för att jämföra systemen var Root Mean Square av Absolute Trajectory Error i meter. Resultaten i detta arbete visar på att DM-VIO är i hög-grad beroende av den visuella signalen som används, och prestandan minskar avsevärt när rörelseoskärpan ökar, oavsett sensorkonfiguration. När enbart en kamera (Monocular) används minskar prestandan från centimeterprecision till diameter. ORB-SLAM3 påverkas inte av rörelseoskärpa och presterar med centimeterprecision för alla sekvenser. Det kan heller inte påvisas någon signifikant skillnad mellan sensorkonfigurationerna. Trots detta kan ett stokastiskt beteende i ORB-SLAM3 noteras, detta kan ha orsakat vissa sekvenser att bete sig avvikande. I helhet, ORB-SLAM3 överträffar DM-VIO på alla sekvenser i det semisyntetiska datasetet som skapats för detta arbete. Koden som använts i detta arbete finns tillgängligt på GitHub https://github.com/cmannila tillsammans med forkade repository för DM-VIO och ORB-SLAM3.

This study investigates the feasibility and performance of SLAM (Simultaneous Localization and Mapping) as a service (SLAM-as-a-Service) for outdoor augmented reality (AR) applications. Given the rapid advancements in AR technology, integrating lightweight AR glasses with real-time SLAM capabilities poses significant challenges, particularly due to the computational demands of SLAM algorithms and the limited hardware capacity of AR devices. This study proposes a scalable SLAM-as-a-Service framework that offloads intensive computational tasks to remote servers, leveraging cloud and edge computing resources. The ORB-SLAM3 algorithm, known for its robustness and real-time processing capabilities, was adapted and implemented in a service-oriented architecture. The framework was evaluated using the EuRoC dataset to benchmark processing speed, accuracy, and round trip time. The results indicate that while the proposed SLAM-as-a-Service model shows promise in handling high computational loads, several obstacles need to be addressed to achieve minimal round trip time and ensure a seamless AR experience. This thesis contributes to the development of scalable and efficient AR solutions by addressing the limitations of on device processing and highlighting the potential of cloud-based services in enhancing the performance and feasibility of AR applications in dynamic outdoor environments.

This thesis analyzes the impact of and implements compensation for rolling shutter distortions in the state-of-the-art feature-based visual SLAM system ORB-SLAM3. The compensation method involves rectifying the detected features, and the evaluation was conducted on the "Rolling-Shutter Visual-Inertial Odometry Dataset" from TUM, which comprises of ten sequences recorded with side-by-side synchronized global and rolling shutter cameras in a single room.  The performance of ORB-SLAM3 on rolling shutter without the implemented rectification algorithms substantially decreased in terms of accuracy and robustness. The global shutter camera achieved centimeter or even sub-centimeter accuracy, while the rolling shutter camera's accuracy could reach the decimeter range in the more challenging sequences. Also, specific individual executions using a rolling shutter camera could not track the trajectory effectively, indicating a degradation in robustness. The effects of rolling shutter in inertial ORB-SLAM3 were even more pronounced with higher trajectory errors and outright failure to track in some sequences. This was the case even though using inertial measurements with the global shutter camera resulted in better accuracy and robustness compared to the non-inertial case.  The rectification algorithms implemented in this thesis yielded significant accuracy increases of up to a 7x relative improvement for the non-inertial case, which turned trajectory errors back to the centimeter scale from the decimeter one for the more challenging sequences. For the inertial case, the rectification scheme was even more crucial. It resulted in better trajectory accuracies, better than the non-inertial case for the less challenging sequences, and made tracking possible for the more challenging ones.

In this thesis, the robustness of deep learning techniques in the field of visual odometry is investigated, with a specific focus on the impact of motion blur. A comparative study is conducted, evaluating the performance of state-of-the-art deep convolutional neural network methods, namely DF-VO and DytanVO, against ORB-SLAM3, a well-established non-deep-learning technique for visual simultaneous localization and mapping. The objective is to quantitatively assess the performance of these models as a function of motion blur. The evaluation is carried out on a custom synthetic dataset, which simulates a camera navigating through a forest environment. The dataset includes trajectories with varying degrees of motion blur, caused by camera translation, and optionally, pitch and yaw rotational noise. The results demonstrate that deep learning-based methods maintained robust performance despite the challenging conditions presented in the test data, while excessive blur lead to tracking failures in the geometric model. This suggests that the ability of deep neural network architectures to automatically learn hierarchical feature representations and capture complex, abstract features may enhance the robustness of deep learning-based visual odometry techniques in challenging conditions, compared to their geometric counterparts.