The spatial frequency doubling illusion (FDI) occurs when the contrast of a low spatial frequency sinusoidal grating is modulated at high temporal frequencies – its apparent spatial frequency increases. Earlier suggestions were that the FDI is generated by a specific class of retinal ganglion cells, which are preferentially lost in the early stages of glaucoma. Based on this linking theory, frequency doubling perimetry (FDP) was developed and several clinical reports confirmed its high efficiency in diagnosing early glaucomatous vision loss. However, this linking theory is not universally accepted and newer suggestions posit that the illusion arises because of temporal frequency related difficulties in temporal phase encoding ability. This thesis psychophysically examines the spatiotemporal characteristics of both the FDI and temporal phase encoding ability with achromatic and equi-luminant (both red-green (RG) and blue-yellow (BY)) gratings at a range of spatiotemporal parameters including those eliciting the FDI. (For complete abstract open document)

Reasoning about temporal and spatial information is a common task in computer science, especially in the field of artificial intelligence. The topic of this thesis is the study of such reasoning from a computational perspective. We study a number of different qualitative point based formalisms for temporal reasoning and provide a complete classification of computational tractability for different time models. We also develop more general methods which can be used for proving tractability and intractability of other relational algebras. Even though most of the thesis pertains to qualitative reasoning the methods employed here can also be used for quantitative reasoning. For instance, we introduce a tractable and useful extension to the quantitative point based formalism STP. This extension gives the algebra an expressibility which subsumes the largest tractable fragment of the augmented interval algebra and has a faster and simpler algorithm for deciding consistency. The use of disjunctions in temporal formalisms is of great interest not only since disjunctions are a key element in different logics but also since the expressibility can be greatly enhanced in this way. If we allow arbitrary disjunctions, the problems under consideration typically become intractable and methods to identify tractable fragments of disjunctive formalisms are therefore useful. One such method is to use the independence property. We present an automatic method for deciding this property for many relational algebras. Furthermore, we show how this concept can not only be used for deciding tractability of sets of relations but also to demonstrate intractability of relations not having this property. Together with other methods for making total classifications of tractability this goes a long way towards easing the task of classifying and understanding relational algebras. The tractable fragments of relational algebras are sometimes not expressive enough to model real-world problems and a backtracking solver is needed. For these cases we identify another property among relations which can be used to aid general backtracking based solvers to finnd solutions faster. / Article I is a revised and extended version of the following three papers: 1. Mathias Broxvall and Peter Jonsson. Towards a Complete Classification of Tractability in Point Algebras for Nonlinear Time. In Proceedings of the 5th International Conference on Principles and Practice of Constraint Programming (CP-99), pp. 129-143, Alexandria, VA, USA, Oct, 1999. 2. Mathias Broxvall and Peter Jonsson. Disjunctive Temporal Reasoning in Partially Ordered Time Structures. In Proceedings of the Seventeenth National Conference on Artificial Intelligence (AAAI-2000), pp. 464-469, Austin, Texas, USA, Aug, 2000. 3. Mathias Broxvall. The Point Algebra for Branching Time Revisited. In Proceedings of the Joint German/Austrian Conference on Artificial Intelligence (KI-2001), pp. 106-121, Vienna, Austria, Sep, 2001. --- Article II is a revised and extended version of the following paper: Mathias Broxvall, Peter Jonsson and Jochen Renz: Refinements and Independence: A Simple Method for Identifying Tractable Disjunctive Constraints. In Proceedings of the 6th International Conference on Principles and Practice of Constraint Programming (CP-2000), pp. 114-127, Singapore, Sep, 2000.

Variations in Leaf Area Index (LAI) can greatly alter output values and patterns of various models that deal with energy flux exchange between the land surface and the atmosphere. Customarily, such models are initiated by LAI estimated from satellite-level Vegetation Indices (VIs) including routinely produced Normalized Difference Vegetation Index (NDVI) products. However, the accuracy from LAI-VI relationships greatly varies due to many factors, including temporal and spatial variations in LAI and a selected VI. In addition, NDVI products derived from various sensors have demonstrated variations in a certain degree on describing temporal and spatial variations in LAI, especially in semi-arid areas. This thesis therefore has three objectives: 1) determine a suitable VI for quantifying LAI temporal variation; 2) improve LAI estimation by considering both temporal and spatial variations in LAI; and 3) evaluate routinely produced NDVI products on monitoring temporal and spatial variations in LAI.<p> The study site was set up in conserved semi-arid mixed grassland in St. Denis, Saskatchewan, Canada. One 600 m - long sampling transect was set up across the rolling typography, and six plots with a size of 40 × 40 m each were randomly designed and each was in a relatively homogenous area. Plant Area Index (PAI, which was validated to obtain LAI), ground hyperspectral reflectance, ground covers (grasses, forbs, standing dead, litter, and bare soil), and soil moisture data were collected over the sampling transect and plots from May through September, 2008. Satellite data used are SPOT 4/5 images and 16-day Moderate Resolution Imaging Spectroradiometer (MODIS) 250m, 1km as well as 10-day SPOT-vegetation (SPOT-VGT) NDVI products from May to October, 2007 and 2008. The results show that NDVI is the most suitable VI for quantifying temporal variation of LAI. LAI estimation is much improved by considering both temporal and spatial variations. Based on the ground reflectance data, the r2 value is increased by 0.05, 0.31, and 0.23 and an averaged relative error is decreased by 1.57, 1.62, and 0.67 in the early, maximum, and late growing season, respectively. MODIS 250m NDVI products are the most useful datasets and MODIS 1km NDVI products are superior to SPOT-VGT 1km composites for monitoring intra-annual spatiotemporal variations in LAI. The proposed LAI estimation approach can be used in other studies to obtain more accurate LAI, and thus this research will be beneficial for grassland modeling.

Variations in Leaf Area Index (LAI) can greatly alter output values and patterns of various models that deal with energy flux exchange between the land surface and the atmosphere. Customarily, such models are initiated by LAI estimated from satellite-level Vegetation Indices (VIs) including routinely produced Normalized Difference Vegetation Index (NDVI) products. However, the accuracy from LAI-VI relationships greatly varies due to many factors, including temporal and spatial variations in LAI and a selected VI. In addition, NDVI products derived from various sensors have demonstrated variations in a certain degree on describing temporal and spatial variations in LAI, especially in semi-arid areas. This thesis therefore has three objectives: 1) determine a suitable VI for quantifying LAI temporal variation; 2) improve LAI estimation by considering both temporal and spatial variations in LAI; and 3) evaluate routinely produced NDVI products on monitoring temporal and spatial variations in LAI.<p> The study site was set up in conserved semi-arid mixed grassland in St. Denis, Saskatchewan, Canada. One 600 m - long sampling transect was set up across the rolling typography, and six plots with a size of 40 × 40 m each were randomly designed and each was in a relatively homogenous area. Plant Area Index (PAI, which was validated to obtain LAI), ground hyperspectral reflectance, ground covers (grasses, forbs, standing dead, litter, and bare soil), and soil moisture data were collected over the sampling transect and plots from May through September, 2008. Satellite data used are SPOT 4/5 images and 16-day Moderate Resolution Imaging Spectroradiometer (MODIS) 250m, 1km as well as 10-day SPOT-vegetation (SPOT-VGT) NDVI products from May to October, 2007 and 2008. The results show that NDVI is the most suitable VI for quantifying temporal variation of LAI. LAI estimation is much improved by considering both temporal and spatial variations. Based on the ground reflectance data, the r2 value is increased by 0.05, 0.31, and 0.23 and an averaged relative error is decreased by 1.57, 1.62, and 0.67 in the early, maximum, and late growing season, respectively. MODIS 250m NDVI products are the most useful datasets and MODIS 1km NDVI products are superior to SPOT-VGT 1km composites for monitoring intra-annual spatiotemporal variations in LAI. The proposed LAI estimation approach can be used in other studies to obtain more accurate LAI, and thus this research will be beneficial for grassland modeling.

Because of the dynamic nature of shrinking and swelling of soils that are classified as Vertisols, partitioning of rainfall into infiltration and runoff in a Vertic watershed is more temporally and spatially unique than in most other watersheds. Hydrology models that account for realistic representation of crack dynamics are rarely used because the spatial and temporal patterns of cracking across a catena and under different land uses are poorly understood. The objectives of the study were to 1) determine if variability in soil cracking on a Vertisol catena, having the same soil and land cover, could be explained by shrink-swell potential of the soil and changes in soil water content; 2) characterize the temporal and spatial variability of the shrinkage of a Vertisol under different land uses; and 3) determine the relationship between specific volume and water content of soils, particularly between saturation and field capacity. The research was conducted in Vertisol catenae of the Houston Black and Heiden soil series. The catenae were located within the USDA-ARS Grassland, Soil and Water Research Laboratory, Riesel Texas. Soil samples were taken to characterize the general properties of the soils. In situ bi-weekly measurements of vertical soil movements and soil water contents were made over a two-year span. Because shrink-swell potential was high at most landscape positions, soil water content was the primary factor driving the spatial and temporal variability of soil shrinking and swelling. The measured relationship between the amount of soil subsidence and water loss generally agreed with what would be theoretically expected. Maximum soil subsidence was 120 mm in the grazed pasture, 75 mm in the native prairie, and 76 mm in the row cropped field. Shrinkage of the whole soil was not equidimensional, and the study generally indicates more horizontal shrinkage than vertical shrinkage. Laboratory analysis showed an appreciable change in volume of soils between saturation and field capacity, suggests a layer of soil layer can subside up to 4% while drying from saturation to field capacity, which indicates the common laboratory measure of shrink swell potential does not capture the complete shrink-swell behavior of soils.

This study aimed at providing a thorough understanding of temporal and spatial variations of VOCs and underlying factors in different microenvironments in two different urban atmospheres, with different degrees of regulatory enforcement. The VOC data were collected in field campaigns conducted in Ankara, Turkey, and Ottawa, Canada over the years 2000-2004. Insight into the sources of VOCs in different urban atmospheres was sought by using three commonly used receptor models namely / Positive Matrix Factorization (PMF), Chemical Mass Balance (CMB) Model and Conventional Factor Analysis (CFA). Motor vehicle related source profiles were developed to use in receptor modeling. Motor vehicles are the most abundant VOC sources with about 60% and 95% contributions to ambient levels in Ankara and Ottawa, respectively. Residential heating (31%) during winter season, biogenic (9%) and architectural coating (12%) emissions during summer season and solvent use (about 12%) emissions are the next abundant VOC sources in Ankara. In addition, a new method to estimate the contribution of sources from wind sectors in urban atmosphere was developed and implemented in this study. The comparison of the results of these two cities demonstrated the influence of control measures on ambient levels and sources of VOCs observed in different urban atmospheres. VOC levels in Ankara exceed EU levels and they are about factor of two higher than that are measured in Ottawa owing to lack of implementation of emission control regulations for VOCs in Ankara compared to well adopted regulations in Ottawa.

The spatial frequency doubling illusion (FDI) occurs when the contrast of a low spatial frequency sinusoidal grating is modulated at high temporal frequencies – its apparent spatial frequency increases. Earlier suggestions were that the FDI is generated by a specific class of retinal ganglion cells, which are preferentially lost in the early stages of glaucoma. Based on this linking theory, frequency doubling perimetry (FDP) was developed and several clinical reports confirmed its high efficiency in diagnosing early glaucomatous vision loss. However, this linking theory is not universally accepted and newer suggestions posit that the illusion arises because of temporal frequency related difficulties in temporal phase encoding ability. This thesis psychophysically examines the spatiotemporal characteristics of both the FDI and temporal phase encoding ability with achromatic and equi-luminant (both red-green (RG) and blue-yellow (BY)) gratings at a range of spatiotemporal parameters including those eliciting the FDI. (For complete abstract open document)

O objetivo do trabalho foi caracterizar a transmissão de dengue nas regiões de São José do Rio Preto e Araçatuba, entre 1990 e 1996. Analisaram-se os Coeficientes de Incidência (CI) de casos confirmados laboratorialmente e autóctones de dengue segundo áreas, sexo e idade. Analisaram-se as curvas de CI mensais para as regiões, sub-regiões e municípios. Realizou-se análise espacial, relacionando-se os CI com variáveis sócio-econômicas e demográficas. Construíram-se curvas de CI semanais de casos suspeitos e confirmados laboratorialmente de dengue para investigar a relação entre elas. Os CI obtidos apontam para o risco de ocorrência de dengue hemorrágica. Verificaram-se maiores CI para as mulheres e para pessoas entre 30 e 59 anos. Nos dois últimos períodos estudados, notaram-se, para as várias sub-regiões, maiores níveis de incidência e aumento do número de meses com transmissão, indicando uma possível tendência de endemização da doença. Verificou-se que foi pelos municípios maiores que a transmissão, em geral, se iniciou e se encerrou, e que neles a freqüência de períodos com transmissão foi maior. Verificou-se também, que, nos municípios com melhores situações de instrução, renda e maiores populações, ocorreu, em maior proporção, CI de valores mais altos, podendo, este fato, estar relacionado com a atração exercida por eles. Estas evidências mostram que um possível fator de risco a ser utilizado, em uma estratificação de medidas de controle, seria o tamanho do município. Foram obtidas relações entre as séries de CI de casos suspeitos e confirmados laboratorialmente de dengue, possibilitando-se estimar as últimas a partir das primeiras. / This study aims to characterize the dengue transmission in th e regions of São José do Rio Preto and Araçatuba, São Paulo, between 1990 and 1996. Incidence Rates (IR) of in laboratory confirmed dengue cases as well as autochthonous dengue cases have been analyzed according to areas, gender and age. Monthly IR curves have been analyzed for the regions, sub-regions and municipal districts. A spatial analysis has been carried out relating the IR with social-economical and demographic variables. Weekly IR curves of both suspected cases and in laboratory confirmed dengue cases have been built in order to investigate the relationship between them. The obtained IR point for a risk of hemorrhagic dengue occurrence. It was noticed a higher risk of dengue occurrence for women and among people between 30 and 59 years. In the last studied periods it was also noticed greater incidence in several sub-areas and increase of month periods with transmission, indicating a possible disease spreading. The transmission, in general, started and finished in large municipal districts, and in those the transmission period was more frequent. It was also observed that, in large municipal districts with better educational level and income variables, IR showing higher levels occured in a larger exrtent than in smaller ones whose indicators of those variables were worst, and the explanation for this can be related with the attraction exercised by the great population centers. These evidences have showed that a possible risk factor to be used in the stratification of the control measures would be the size of the municipal district. It was obtained the relationship between the IR series of suspected dengue cases and in laboratory confirmed cases enabling us to measure the last ones starting from the first ones.

Indiana University-Purdue University Indianapolis (IUPUI) / In traffic scene perception for autonomous vehicles, driving videos are available from in-car sensors such as camera and LiDAR for road detection and collision avoidance. There are some existing challenges in computer vision tasks for video processing, including object detection and tracking, semantic segmentation, etc. First, due to that consecutive video frames have a large data redundancy, traditional spatial-to-temporal approach inherently demands huge computational resource. Second, in many real-time scenarios, targets move continuously in the view as data streamed in. To achieve prompt response with minimum latency, an online model to process the streaming data in shift-mode is necessary. Third, in addition to shape-based recognition in spatial space, motion detection also replies on the inherent temporal continuity in videos. While current works either lack long-term memory for reference or consume a huge amount of computation. The purpose of this work is to achieve strongly temporal-associated sensing results in real-time with minimum memory, which is continually embedded to a pragmatic framework for speed and path planning. It takes a temporal-to-spatial approach to cope with fast moving vehicles in autonomous navigation. It utilizes compact road profiles (RP) and motion profiles (MP) to identify path regions and dynamic objects, which drastically reduces video data to a lower dimension and increases sensing rate. Specifically, we sample one-pixel line at each video frame, the temporal congregation of lines from consecutive frames forms a road profile image; while motion profile consists of the average lines by sampling one-belt pixels at each frame. By applying the dense temporal resolution to compensate the sparse spatial resolution, this method reduces 3D streaming data into 2D image layout. Based on RP and MP under various weather conditions, there have three main tasks being conducted to contribute the knowledge domain in perception and planning for autonomous driving. The first application is semantic segmentation of temporal-to-spatial streaming scenes, including recognition of road and roadside, driving events, objects in static or motion. Since the main vision sensing tasks for autonomous driving are identifying road area to follow and locating traffic to avoid collision, this work tackles this problem by using semantic segmentation upon road and motion profiles. Though one-pixel line may not contain sufficient spatial information of road and objects, the consecutive collection of lines as a temporal-spatial image provides intrinsic spatial layout because of the continuous observation and smooth vehicle motion. Moreover, by capturing the trajectory of pedestrians upon their moving legs in motion profile, we can robustly distinguish pedestrian in motion against smooth background. The experimental results of streaming data collected from various sensors including camera and LiDAR demonstrate that, in the reduced temporal-to-spatial space, an effective recognition of driving scene can be learned through Semantic Segmentation. The second contribution of this work is that it accommodates standard semantic segmentation to sequential semantic segmentation network (SE3), which is implemented as a new benchmark for image and video segmentation. As most state-of-the-art methods are greedy for accuracy by designing complex structures at expense of memory use, which makes trained models heavily depend on GPUs and thus not applicable to real-time inference. Without accuracy loss, this work enables image segmentation at the minimum memory. Specifically, instead of predicting for image patch, SE3 generates output along with line scanning. By pinpointing the memory associated with the input line at each neural layer in the network, it preserves the same receptive field as patch size but saved the computation in the overlapped regions during network shifting. Generally, SE3 applies to most of the current backbone models in image segmentation, and furthers the inference by fusing temporal information without increasing computation complexity for video semantic segmentation. Thus, it achieves 3D association over long-range while under the computation of 2D setting. This will facilitate inference of semantic segmentation on light-weighted devices. The third application is speed and path planning based on the sensing results from naturalistic driving videos. To avoid collision in a close range and navigate a vehicle in middle and far ranges, several RP/MPs are scanned continuously from different depths for vehicle path planning. The semantic segmentation of RP/MP is further extended to multi-depths for path and speed planning according to the sensed headway and lane position. We conduct experiments on profiles of different sensing depths and build up a smoothly planning framework according to their them. We also build an initial dataset of road and motion profiles with semantic labels from long HD driving videos. The dataset is published as additional contribution to the future work in computer vision and autonomous driving.

Coastal foredunes are natural aeolian bedforms located landward of the backshore and which interact continuously with the beach. Traditionally, coastal dunes have been associated with onshore winds, however they can be found under more complex wind regimes where offshore winds are common such as the UK East coast, Northern Ireland and New Zealand. This research investigates the ways in which foredune-beach interactions occur under a complex wind regime at a range of overlapping temporal and spatial scales and is innovative in that it explicitly links small-scale processes and morphodynamic behaviour to large scale and long-term dynamics. The study area is the north Lincolnshire coast, East England. Detailed observations of airflow at three locations under varying wind regimes revealed considerable spatial variations in wind velocity and direction, however it was possible to determine a general model of how foredune topography deflected and modified airflow and the resultant geomorphological implications (i.e. erosion and deposition). During direct offshore and onshore winds, airflow remained attached and undeflected; and distinct zones of flow deceleration and acceleration could be identified. During oblique winds airflow was deflected to become more parallel to the dune crest. The field sites used are characterized by a seasonal erosion/accretion cycle and a series of increasingly complex models was developed and tested to determine whether it was possible to predict sand volume changes in the foredune-beach system based on a limited number of variables. The model predictions were tested against detailed digital terrain models at a seasonal timescale. The model prediction that best matched the observed (surveyed) sand volume changes included wind speed, direction, grain size, fetch effect controlled by beach inundation and angle of wind approach was accurate to within ±10% for 18 out of 48 tests at the seasonal scale and 6 out of 12 tests over periods of >5 years. A key variable influencing foredune-beach sand volume is the magnitude and frequency of storm surge events and this was not factored in to the model, but may explain the model-observation mismatch over the medium-term on two occasions. Over the past 120 years historical maps and aerial photographs indicate long-term foredune accretion of approximately 2 m year-1 at the three study sites (1891-2010). At this timescale, rates of coastal foredune accretion reflect the low occurrence of severe storm surges and suggest rapid post-storm recovery. The morphological response of the foredune-beach morphology is considered to be a combination of controlling and forcing factors. Process-responses within the system, associated with nearshore interactions and sediment transfer from the littoral drift, are compiled into a multi-scale morphodynamic model. Important to match appropriate dataset to scale of research question or management plan being explored. In the case of management, long-term records of past activity are necessary to predict the future but also to understand natural responses of system to short-term impact such as storm surge.