In this thesis a new image-based texturing method has been developed. This new method allows users to synthesize tiled textures that can be mapped to any quadrilateral mesh without discontinuity or singularity. An interface has been developed that allows user control over out put textures. Three methods have been included in the interface to create a periodic looking texture for 3D models and two methods have been developed to create wallpaper images (repeating textures on a 2D surface). Using these texturing methods, texturing problems are simplified, and more time can be spent solving artistic problems.

We present a novel technique for capturing spatially or temporally resolved light probe sequences, and using them for image based lighting. For this purpose we have designed and built a real-time light probe, a catadioptric imaging system that can capture the full dynamic range of the lighting incident at each point in space at video frame rates, while being moved through a scene. The real-time light probe uses a digital imaging system which we have programmed to capture high quality, photometrically accurate color images of 512×512 pixels with a dynamic range of 10000000:1 at 25 frames per second. By tracking the position and orientation of the light probe, it is possible to transform each light probe into a common frame of reference in world coordinates, and map each point and direction in space along the path of motion to a particular frame and pixel in the light probe sequence. We demonstrate our technique by rendering synthetic objects illuminated by complex real world lighting, first by using traditional image based lighting methods and temporally varying light probe illumination, and second an extension to handle spatially varying lighting conditions across large objects and object motion along an extended path.

Detta arbete har gjorts i samarbete med Rikspolisstyrelsen för att ta fram en metod för hur modern medieteknik kan användas för att skapa en ”virtuell brottsplats”. Syftet är att arbetet ska leda till ett förslag till en metod som lämpar sig för att integrera i polisens brottsplatsundersökningar och rättsliga processer, med beaktande av de speciella krav som ställs. Arbetet innehåller två huvuddelar där den första delens utgångspunkt är vad som går att göra med utrustning och teknik som redan finns tillgänglig och den andra delen hur det skulle kunna utvecklas vidare. Till första delen har ett förslag på en metod som kan användas för att utnyttja panoramatekniken, tagits fram. Därför har det också genomförts utvärderingar och tester på befintliga programvaror för att utröna vad som passar syftet bäst. För den andra delen togs en egen lösning fram och implementerades i OpenGL/C++. Denna lösning baseras på laserskanningsdata. Resultatet av denna del är inte en färdig metod som kan börja användas direkt utan mer ett exempel på hur panoramatekniken kan användas till något mer än att bara visa hur en plats ser ut. För att knyta samman projektet med verkligheten har båda dessa delar tillämpats på flera riktiga fall. En slutsats som kan dras av arbetet är att visualiseringar av denna typ är väldigt användbara och till fördel för utredare och åklagare. Det finns mycket kvar att undersöka men det är ingen tvekan om att den här typen av teknik är användbar för detta syfte.

This thesis presents a web-based approach to lighting three-dimensional geometry in a virtual scene. The use of High Dynamic Range (HDR) images for the lighting model makes it possible to convey a greater sense of photorealism than can be provided with a conventional computer generated three-point lighting setup. The use of QuickTime ™ Object Virtual Reality to display the three-dimensional geometry offers a sophisticated user experience and a convenient method for viewing virtual objects over the web. With this work, I generate original High Dynamic Range images for the purpose of image-based lighting and use the QuickTime ™ Object Virtual Reality framework to creatively alter the paradigm of object VR for use in object lighting. The result is two scenarios: one that allows for the virtual manipulation of an object within a lit scene, and another with the virtual manipulation of light around a static object. Future work might include the animation of High Dynamic Range image-based lighting, with emphasis on such features as depth of field and glare generation.

In this thesis a new image-based texturing method has been developed. This new method allows users to synthesize tiled textures that can be mapped to any quadrilateral mesh without discontinuity or singularity. An interface has been developed that allows user control over out put textures. Three methods have been included in the interface to create a periodic looking texture for 3D models and two methods have been developed to create wallpaper images (repeating textures on a 2D surface). Using these texturing methods, texturing problems are simplified, and more time can be spent solving artistic problems.

Due to the rapid development of 3D display market, the protection and authentication of the intellectual property rights of 3D multimedia has become an essential concern. As a consequence, the digital watermarking for 3D image and video is attracting considerable attention. The depth-image-based rendering (DIBR) technique has been playing a critical role in 3D contents representation because of its numerous advantages. A good digital watermarking algorithm should be robust to various possible attacks, including geometric distortions and compressions. And di erent from ordinary 2D digital watermarking, there are more speci c requirements for 3D watermarking, especially for DIBR 3D image watermarking. Not only the center view, but also the virtual left and right views can be illegally distributed. Therefore, the embedded watermark information should be accurately extracted from these three views individually for content authentication, even under attacks. In this thesis, we focus on the research of digital watermarking and watermarking based quality evaluation for DIBR 3D images. We first present a 2D image and video watermarking method based on contourlet transform, which is then extended to a robust contourlet-based watermarking algorithm for DIBR 3D images. The watermark is embedded into the center view by quantizing certain contourlet coe cients. The virtual left and right views are synthesized from the watermarked center view and the corresponding depth map. One advantage of our algorithm is its simplicity and practicality. However, the performance on watermark extraction needs to be further improved. As an improvement, a blind watermarking algorithm for DIBR 3D images based on feature regions and ridgelet transform is proposed. The watermarked view has good perceptual quality under both the objective and subjective image quality measures. Compared with other related and state-of-the-art methods, the proposed algorithm shows superiority in terms of watermark extraction and robustness to various attacks. Furthermore, as one of the most promising techniques for quality evaluation, a watermarking based quality evaluation scheme is developed for DIBR 3D images. The qualities of the watermarked center view and the synthesized left and right views under distortions can be estimated by examining the degradation of corresponding extracted watermarks. The simulation results demonstrate that our scheme has good performance of quality evaluation for DIBR 3D images under the attacks.

With numerous distinct advantages, quadrotors have found a wide range of applications, such as structural inspection, traffic control, search and rescue, agricultural surveillance, etc. To better serve applications in cluttered environment, quadrotors are further equipped with vision sensors to enhance their state sensing and environment perception capabilities. Moreover, visual information can also be used to guide the motion control of the quadrotor. This is referred to as visual servoing of quadrotor. In this thesis, we identify the challenging problems arising in the area of visual servoing of the quadrotor and propose effective control strategies to address these issues. The control objective considered in this thesis is to regulate the relative pose of the quadrotor to a ground target using a limited number of sensors, e.g., a monocular camera and an inertia measurement unit. The camera is attached underneath the center of the quadrotor and facing down. The ground target is a planar object consisting of multiple points. The image features are selected as image moments defined in a ``virtual image plane". These image features offer an image kinematics that is independent of the tilt motion of the quadrotor. This independence enables the separation of the high level visual servoing controller design from the low level attitude tracking control. A high-gain observer-based model predictive control (MPC) scheme is proposed in this thesis to address the image-based visual servoing of the quadrotor. The high-gain observer is designed to estimate the linear velocity of the quadrotor which is part of the system states. Due to a limited number of sensors on board, the linear velocity information is not directly measurable. The high-gain observer provides the estimates of the linear velocity and delivers them to the model predictive controller. On the other hand, the model predictive controller generates the desired thrust force and yaw rate to regulate the pose of the quadrotor relative to the ground target. By using the MPC controller, the tilt motion of the quadrotor can be effectively bounded so that the scene of the ground target is well maintained in the field of view of the camera. This requirement is referred to as visibility constraint. The satisfaction of visibility constraint is a prerequisite of visual servoing of the quadrotor. Simulation and experimental studies are performed to verify the effectiveness of the proposed control strategies. Moreover, image processing algorithms are developed to extract the image features from the captured images, as required by the experimental implementation. / Graduate / 2020-12-11

<p>Pavement macrotexture contributes greatly to road surface friction, which in turn plays a significant role in reducing road incidents. Conventional methods for macrotexture measurement techniques (e.g., the sand patch method, the outflow method, and laser measuring) are either expensive, time-consuming, or of poor repeatability. This thesis aims to develop and evaluate affordable and convenient alternative approaches to determine pavement macrotexture. The proposed solution is based on multi-view smartphone images collected in situ over the pavement. Computer vision techniques are then applied to create high resolution three-dimensional (3D) models of the pavement. The thesis develops the analytics to determine two primary macrotexture metrics: mean profile depth and aggregation loss. Experiments with 790 images over 25 spots of three State Roads and 6 spots of the INDOT test site demonstrated that the image-based method can yield reliable results comparable to conventional laser texture scanner results. Moreover, based on experiments with 280 images over 7 sample plates with different aggregate loss percentage, the newly developed analytics were proven to enable estimation of the aggregation loss, which is largely compromised in the laser scanning technique and conventional MPD calculation approach. The root mean square height based on the captured images was verified in this thesis as a more comprehensive metric for macrotexture evaluation. It is expected that the developed approach and analytics can be adopted for practical use at a large scale. </p>

Preclinical research for developing new drugs is a long and expensive procedure. Experiments relying on image acquisition and analysis tend to be low throughput and use reporter systems that may influence the studied cells. With image-based assays focusing on extracting qualitative information from microscopic images of mammalian cells, more cost-efficient and high-throughput analyses are possible. Furthermore, studying cell morphology has proven to be a good indicator of cell phenotype. Using hand-crafted feature descriptors based on cell morphology, label-free quantification of cell apoptosis has been achieved. These hand-crafted descriptors are based on cell characteristics translated to quantifiable metrics, but risk being biased towards easily observable features and therefore miss subtle ones.         This project proposes an alternative approach by generating a latent representation of cell features using deep learning models and aims to find if they can compete with pre-defined hand-crafted representations in classifying live or dead cells. For this purpose, three deep learning models are implemented, one autoencoder and two variational-autoencoder. We develop a core architecture shared between the models based on a convolutional neural network using a latent space with 16 dimensions. We then train the models to recreate single-cell images of SKOV3 ovarian cancer cells. The latent representation was extracted at specific checkpoints during training and later used for training a logistic regression classifier. Finally, comparing classification accuracy between the hand-crafted feature representations and generated representation was made with novel cell images. The generated representations show a slight but consistent increase in classification accuracy, up to 4.9 percent points, even without capturing all morphological details in the recreation. Thus, we conclude that it is possible for generated representations to outperform hand-crafted feature descriptors in live or dead cell classification.

As existing bridge structures age, they are susceptible to the effects of deterioration, damage and other deleterious processes. These effects hamper the capacity and efficiency of transportation networks and adversely impact local, regional and national economic growth. As a result, bridge authorities and other professionals have become more sensitive to maintenance issues related to this aging infrastructure. While highway bridge condition have been monitored by visual inspection, non-destructive evaluation (NDE) technologies have also been developing and they are expected to be utilized for effective management of highway bridges or other civil infrastructure systems. Efficient use of these technologies saves time spent or bridge inspections, and also helps the bridge authorities for management decision-making. One of the NDE technologies is the image-based technology. In this thesis research, image-based technologies using high resolution digital images (HRDI) and infrared thermography image (IRTI) are introduced, described and implemented. First, a review of the mechanisms of these technologies is presented. Due to the specific engineering utilization and recent technological development, there is a need to validate effectiveness of HRDI and IRTI for their practical use for engineering purpose. For this reason, a pilot project using these technologies was conducted at an in-service bridge and a parking structure with the support of Florida Department of Transportation District 5 and the results are presented in this thesis. Secondly, in order to explore and enhance the usability of infrared thermography technology (IRTI), experiments on campus and on another bridge were conducted to determine the best time to test bridges and the sensitivity of IRTI to delamination volume. Since the iv accuracy of damage detection using infrared thermography technology is greatly affected by daily temperature variation, it is quite important to estimate an appropriate duration for infrared thermography inspection prior to the inspection. However, in current practice, the way to estimate the duration is to monitor the temperature of the concrete surface. Since the temperature varies depending on the area or region, there is a need to visit the bridge before the actual test and monitor the temperature variation. This requires additional visits to the bridge site and also access to the bridge for measuring concrete temperature. Sometimes, this can be a practical issue. In this research, in order to estimate an appropriate duration without visiting bridges, a practical method is explored by monitoring and analyzing variation of concrete surface temperature at one location and projected to another location by also incorporating other factors that affect the concrete temperature, such as air temperature and humidity. For this analysis, specially-designed concrete plates of a few types of thickness and shapes are used and the regression analysis is employed to establish a relationship between environmental effects and temperature variation between two different sites. The results have been promising for this research study and it is shown that HRDI and IRTI are excellent technologies for assessing concrete structures in a very practical manner.