This dissertation focuses on the investigation of human perception of stereoscopic 3D image quality and the development of automatic stereoscopic 3D image quality assessment frameworks. In order to assess human perception of visual quality, a human study was conducted and interactions between image quality, depth quality, visual comfort, and 3D viewing quality were inferred. The results indicate that the overall 3D viewing quality can be well predicted from only image quality and depth quality. Between image and depth quality, image quality seems to be the main factor that enables accurate prediction of overall 3D viewing quality. Two other human studies were conducted to study the effect of masking on stereoscopic distortions. Binocular suppression was observed in the stereo images which were distorted by blur, JPEG compression, or JPEG2K compression, however, no such suppression was observed for stereo images distorted by white noise. Further, a facilitation effect was also observed against disparity variation for blur and JPEG2K distorted stereo images while no depth masking effect was observed. Based on these results, I proposed an automatic full-reference (FR) 3D quality assessment framework. In this framework, I used Gabor filterbank responses to model stimulus strength and then synthesize a Cyclopean image from a stereo image pair. Because the quality of this synthesized view is similar to that of a Cyclopean image, which the human visual system recreates from the stereoscopic stimuli, performing the task of 3D quality assessment on synthesized views can deliver better performance. I verified the performance of this FR framework on the LIVE 3D Image Quality Database and the results indicate that applying the proposed framework improves the performance of FR 2D quality assessment algorithms when applied to stereo 3D images. Further, I proposed a no-reference (NR) 3D quality assessment (QA) algorithm based on natural scene statistics in both the spatial and the depth domain. Experiments indicate that the proposed NR algorithm outperforms all 2D FR QA algorithms and most 3D FR QA models in predicting 3D quality of stereo images. Finally, a fourth subjective study was conducted to understand depth quality when stereo content is free from visual discomfort. The result suggests that human perception of depth quality is correlated with the content of the stereo image and the stereoacuity function of human visual system. / text

Στη διπλωματική αυτή εργασία αναλύεται η εκτίμηση του βάθους μίας άκαμπτης σκηνής από κάμερα τοποθετημένη σε αυτοκίνητο που κινείται. Στο κεφάλαιο 1 γίνεται μία εισαγωγή στον τομέα της Υπολογιστικής Όρασης και δίνονται μερικά παραδείγματα εφαρμογών της. Στο κεφάλαιο 2 περιγράφονται βασικές αρχές της προβολικής γεωμετρίας που χρησιμοποιείται ως μαθηματικό υπόβαθρο για τα επόμενα κεφάλαια. Στο κεφάλαιο 3 γίνεται λόγος για το θεωρητικό μοντέλο της κάμερας, των παραμέτρων της και των παραμορφώσεων που υπεισέρχονται στο μοντέλο αυτό. Στο κεφάλαιο 4 αναφέρεται η διαδικασία βαθμονόμησης της κάμερας, μαζί με την υλοποίησή της. Στο κεφάλαιο 5 παρουσιάζονται γενικές κατηγορίες των στερεοσκοπικών αλγορίθμων που χρησιμοποιούνται, καθώς και τα κατάλληλα μέτρα ομοιότητάς τους. Στο κεφάλαιο 6 γίνεται αναφορά στον ανιχνευτή γωνιών Harris και γίνεται η εφαρμογή του τόσο ως προς την ανίχνευση των γωνιών, όσο και ως προς την αντιστοίχιση των 2 εικόνων. Στο κεφάλαιο 7 αναλύεται η θεωρία του αλγόριθμου SIFT και δίνεται ένα παράδειγμα ανίχνευσης και αντιστοίχισης χαρακτηριστικών. Στο κεφάλαιο 8 επισημαίνονται οι βασικές αρχές της επιπολικής γεωμετρίας, καθώς η σημασία της διόρθωσης των εικόνων. Στο κεφάλαιο 9 αναφέρεται η συνολική διαδικασία που ακολουθήθηκε, μαζί με την περιγραφή και την υλοποίηση των μεθόδων εκτίμησης βάθους που χρησιμοποιήθηκαν. / The current master’s thesis analyzes the depth estimation of a rigid scene from a camera attached to a moving vehicle. The first chapter gives an introduction to the field of Computer Vision and provides some examples of its applications. The second chapter describes basic principles of projective geometry that are being used as mathematical background for the next chapters. The third chapter refers to the theoretical modeling of a camera, along with its parameters and the distortions that appear in this model. The forth chapter deals with the camera calibration procedure, along with its implementation. Chapter five presents general categories of stereoscopic algorithms, along with their similarity measures. Chapter six talks about Harris corner detector and its implementation in detecting corners and in the matching process as well. Chapter 7 analyzes the SIFT algorithm theory and gives an example of detecting and matching features. Chapter 8 highlights basic principles of epipolar geometry and stresses out the importance of image rectification. Chapter nine presents the procedure that has been followed, along with the description and implementation of the depth estimation methods that have been used.

No description available.

No description available.

No description available.

No description available.

In this thesis, an innovative algorithm for improving the accuracy of variational space-time stereoscopic reconstruction of ocean surfaces is presented. The space-time reconstruction method, developed based on stereo computer vision principles and variational optimization theory, takes videos captured by synchronized cameras as inputs and produces the shape and superficial pattern of an overlapped region of interest as outputs. These outputs are designed to be the minimizers of the variational optimization framework and are dependent on the estimation of the camera parameters. Therefore, from the perspective of computer vision, the proposed algorithm adjusts the estimation of camera parameters to lower the disagreement between the reconstruction and 2-D camera recordings. From a mathematical perspective, since the minimizers of the variational framework are determined by a set of partial differential equations (PDEs), the algorithm modifies the coefficients of the PDEs based on the current numerical solutions to reduce the minimum of the optimization framework. Our algorithm increases the tolerance to the errors of camera parameters, so the joint operations of our algorithm and the variational reconstruction method can generate accurate space-time models even using videos captured by perturbed cameras as input. This breakthrough prompts the realization of ocean surface reconstruction using videos filmed by remotely-controlled helicopters in the future. A number of techniques, technical or theoretical, are explored to fulfill the development and implementation of the algorithm and relative computation issues. The effectiveness of the proposed algorithm is validated through the statistics applied to real ocean surface reconstructions of data collected from an offshore platform at the Crimean Peninsula, the Black Sea. Moreover, synthetic data generated using computer graphics are customized to simulate various situations that are not recorded in the Crimea dataset for the demonstration of the algorithm.

Autonomous Driving or self driving vehicles are concepts of vehicles knowing their environment and making driving manoeuvres without instructions from a driver. The concepts have been around for decades but has improved significantly in the last years since research in this area has made significant progress. Benefits of autonomous driving include the possibility to decrease the number of accidents in traffic and thereby saving lives. A major challenge in autonomous driving is to acquire 3D information and relations between all objects in surrounding traffic. This is referred to as \textit{spatial perception}. Stereo camera systems have become a central sensor module for advanced driver assistance systems and autonomous driving. For object detection and measurements at large distances stereo vision encounter difficulties. This includes objects being small, having low contrast and the presence of image noise. Having an accurate perception of the environment at large distances is however of high interest for many applications, especially autonomous driving. This thesis proposes a method which tries to increase the range to where generic objects are first detected using a given stereo camera setup. Objects are represented by planes in 3D space. The input image is segmented into the various objects and the 3D plane parameters are estimated jointly. The 3D plane parameters are estimated directly from the stereo image pairs. In particular, this thesis investigates methods to introduce geometric constraints to the segmentation or labeling task, i.e assigning each considered pixel in the image to a plane. The methods provided in this thesis show that despite the difficulties at large distances it is possible to exploit planar primitives in 3D space for obstacle detection at distances where other methods fail. / En autonom bil innebär att bilen har en uppfattning om sin omgivning och kan utifran det ta beslut angående hur bilen ska manövreras. Konceptet med självkörande bilar har existerat i årtionden men har utvecklats snabbt senaste åren sedan billigare datorkraft finns lättare tillgänglig. Fördelar med autonomiska bilar innebär bland annat att antalet olyckor i trafiken minskas och därmed liv räddas. En av de största utmaningarna med autonoma bilar är att få 3D information och relationer mellan objekt som finns i den omgivande trafikmiljön. Detta kallas för spatial perception och innebär att detektera alla objekt och tilldela en korrekt postition till dem. Stereo kamerasystem har fått en central roll för avancerade förarsystem och autonoma bilar. För detektion av objekt på stora avstånd träffar stereo system på svårigheter. Detta inkluderar väldigt små objekt, låg kontrast och närvaron av brus i bilden. Att ha en ackurativ perception på stora avstånd är dock vitalt för många applikationer, inte minst autonoma bilar. Den här rapporten föreslar en metod som försöker öka avståndet till där objekt först upptäcks. Objekt representeras av plan i 3D rymden. Bilder givna från stereo par segmenteras i olika object och plan parametrar estimeras samtidigt. Planens parametrar estimeras direkt från stereo bild paren. Den här rapporten utreder metoder att introducera gemoetriska begränsningar att använda vid segmenteringsuppgiften. Metoderna som presenteras i denna rapport visar att trots den höga närvaron av brus på stora avstånd är det möjligt att estimera geometriska objekt som är starka nog att möjliggöra detektion av objekt på ett avstand där andra metoder misslyckas.

This thesis deals with the design and implementation of a visually guided robotic assembly system. Stereo imaging, three dimensional location extraction and object recognition will be the features of this system. This thesis study considers a system utilizing an eye-in-hand configuration. The system involves a stereo rig mounted on the end effector of a six-DOF ABB IRB-2000 industrial robot. The robot is controlled by a vision system, which uses open-loop control principles. The goal of the system is to assemble basic geometric primitives into their respective templates. Recognition

Micropalaeontology, a discipline that contributes to climate research and hydrocarbon exploration, is driven by the taxonomic analysis of huge volumes of microfossils. Unfortunately, this repetitive analysis is a serious bottleneck to progress because it depends on the scarce time of experts. These issues propel research into computerized taxonomic analysis, including a promising new approach called computer-aided microfossil identification. However, the existing computer-aided system relies on image-based representations, which severely limits its ability to discriminate specimens. These limitations motivate using computer vision to support richer video and shape-based representations, which is the focus of this thesis. An important contribution is a scheme to localize, capture, and extract video and shape-based representations from large microfossil batches. These representations encapsulate information across multiple lighting conditions. In addition, the thesis describes a method based on photometric stereo to correct misalignments in images of the same object illuminated from different directions. Not only does this correction benefit the application at hand, but it can also benefit a variety of other applications. The thesis also introduces a visual-surface reconstruction method based on maximum likelihood estimation, which constructs usable depth maps even from extraordinarily noisy images. State of the art methods lack this capability. By freeing classification from the bounds imposed by images, these contributions significantly advance computerized microfossil identification toward the ultimate goal of a practical and reliable tool for high-throughput taxonomic analysis. / Digital Signals and Image Processing