Spatial Analysis of North Central Texas Traffic Fatalities 2001-2006

Spatial Analysis of North Central Texas Traffic Fatalities 2001-2006

Date: December 2010
Creator: Rafferty, Paula S.
Description: A traditional two dimensional (planar) statistical analysis was used to identify the clustering types of North Central Texas traffic fatalities occurring in 2001-2006. Over 3,700 crash locations clustered in ways that were unlike other researched regions. A two dimensional (x and y coordinates) space was manipulated to mimic a one dimensional network to identify the tightest clustering of fatalities in the nearly 400,000 crashes reported from state agencies from 2003-2006. The roadway design was found to significantly affect crash location. A one dimensional (linear) network analysis was then used to measure the statistically significant clustering of flow variables of after dark crashes and daylight crashes. Flow variables were determined to significantly affect crash location after dark. The linear and planar results were compared and the one dimensional, linear analysis was found to be more accurate because it did not over detect the clustering of events on a network.
Contributing Partner: UNT Libraries
Comparison of IKONOS Derived Vegetation Index and LiDAR Derived Canopy Height Model for Grassland Management.

Comparison of IKONOS Derived Vegetation Index and LiDAR Derived Canopy Height Model for Grassland Management.

Date: December 2009
Creator: Parker, Gary
Description: Forest encroachment is understood to be the main reason for prairie grassland decline across the United States. In Texas and Oklahoma, juniper has been highlighted as particularly opportunistic. This study assesses the usefulness of three remote sensing techniques to aid in locating the areas of juniper encroachment for the LBJ Grasslands in Decatur, Texas. An object based classification was performed in eCognition and final accuracy assessments placed the overall accuracy at 94%, a significant improvement over traditional pixel based methods. Image biomass was estimated using normalized difference vegetation index (NDVI) for 1 meter resolution IKONOS winter images. A high correlation between the sum of NDVI for tree objects and field tree biomass was determined where R = 0.72, suggesting NDVI sum of a tree area is plausible. However, issues with NDVI saturation and regression produced unrealistically high biomass estimates for large NDVI. Canopy height model (CHM) derived from 3-5m LiDAR data did not perform as well. LiDAR typically used for digital elevation model (DEM) production was acquired for the CHM and produced correlations of R = 0.26. This suggests an inability for this particular dataset to identify juniper trees. When points that registered a tree height where correlated with field ...
Contributing Partner: UNT Libraries
County Level Population Estimation Using Knowledge-Based Image Classification and Regression Models

County Level Population Estimation Using Knowledge-Based Image Classification and Regression Models

Date: August 2010
Creator: Nepali, Anjeev
Description: This paper presents methods and results of county-level population estimation using Landsat Thematic Mapper (TM) images of Denton County and Collin County in Texas. Landsat TM images acquired in March 2000 were classified into residential and non-residential classes using maximum likelihood classification and knowledge-based classification methods. Accuracy assessment results from the classified image produced using knowledge-based classification and traditional supervised classification (maximum likelihood classification) methods suggest that knowledge-based classification is more effective than traditional supervised classification methods. Furthermore, using randomly selected samples of census block groups, ordinary least squares (OLS) and geographically weighted regression (GWR) models were created for total population estimation. The overall accuracy of the models is over 96% at the county level. The results also suggest that underestimation normally occurs in block groups with high population density, whereas overestimation occurs in block groups with low population density.
Contributing Partner: UNT Libraries
Quantitative Comparison of Lidar Data and User-generated Three-dimensional Building Models From Google Building Maker

Quantitative Comparison of Lidar Data and User-generated Three-dimensional Building Models From Google Building Maker

Date: August 2012
Creator: Liu, Yang
Description: Volunteered geographic information (VGI) has received increased attention as a new paradigm for geographic information production, while light detection and ranging (LiDAR) data is widely applied to many fields. This study quantitatively compares LiDAR data and user-generated 3D building models created using Google Building Maker, and investigate the potential applications of the quantitative measures in support of rapid disaster damage assessment. User-generated 3D building models from Google Building Maker are compared with LiDAR-derived building models using 3D shape signatures. Eighteen 3D building models are created in Fremont, California using the Google Building Maker, and six shape functions (distance, angle, area, volume, slope, and aspect) are applied to the 18 LiDAR-derived building models and user-generated ones. A special case regarding the comparison between LiDAR data and building models with indented walls is also discussed. Based on the results, several conclusions are drawn, and limitations that require further study are also discussed.
Contributing Partner: UNT Libraries
High Resolution Satellite Images and LiDAR Data for Small-Area Building Extraction and Population Estimation

High Resolution Satellite Images and LiDAR Data for Small-Area Building Extraction and Population Estimation

Date: December 2009
Creator: Ramesh, Sathya
Description: Population estimation in inter-censual years has many important applications. In this research, high-resolution pan-sharpened IKONOS image, LiDAR data, and parcel data are used to estimate small-area population in the eastern part of the city of Denton, Texas. Residential buildings are extracted through object-based classification techniques supported by shape indices and spectral signatures. Three population indicators -building count, building volume and building area at block level are derived using spatial joining and zonal statistics in GIS. Linear regression and geographically weighted regression (GWR) models generated using the three variables and the census data are used to estimate population at the census block level. The maximum total estimation accuracy that can be attained by the models is 94.21%. Accuracy assessments suggest that the GWR models outperformed linear regression models due to their better handling of spatial heterogeneity. Models generated from building volume and area gave better results. The models have lower accuracy in both densely populated census blocks and sparsely populated census blocks, which could be partly attributed to the lower accuracy of the LiDAR data used.
Contributing Partner: UNT Libraries
An Investigation of the Relationship between HIV and Prison Facilities in Texas: The Geographic Variation and Vulnerable Neighborhood Characteristics

An Investigation of the Relationship between HIV and Prison Facilities in Texas: The Geographic Variation and Vulnerable Neighborhood Characteristics

Date: August 2011
Creator: Kutch, Libbey
Description: Previous research suggests that prisons may be fueling the spread of HIV infection in the general population. In 2005, the HIV rate was more than 2.5 times higher in US prison populations. Environmental factors in prisons such as illicit drug use and unprotected sexual activities can be conducive for HIV transmission. Because the vast majority of prison inmates are incarcerated for less than three years, transmission of HIV between prison inmates and members of the general population may occur at a high rate. The environment in which an individual lives and the entities that comprise it affect the health of that person. Thus the location of prisons within communities, as well as socio-demographic characteristics may influence the geography of HIV infection. HIV surveillance data, obtained from the Texas Department of State Health Services, were used to investigate the relationship between the location of prison units in Texas and HIV infection rates in the surrounding zip codes. The results suggest that HIV prevalence rates are higher among geographic areas in close proximity to a prison unit. With continued behavioral risks and low treatment adherence rates among individuals infected with HIV, there is a possibility of increased HIV prevalence. Vulnerable places, locations ...
Contributing Partner: UNT Libraries
A Multiscalar Analysis of Buruli Ulcer in Ghana: Environmental and Behavioral Factors in Disease Prevalence

A Multiscalar Analysis of Buruli Ulcer in Ghana: Environmental and Behavioral Factors in Disease Prevalence

Date: May 2012
Creator: Ferring, David
Description: Buruli ulcer (BU), an infectious disease caused by Mycobacterium ulcerans, is the third most common mycobacterial disease after leprosy and tuberculosis and a WHO-defined neglected tropical disease. Despite years of research, the mode of transmission of BU remains unknown. This master’s thesis provides an integrated spatial analysis of disease dynamics in Ghana, West Africa, an area of comparatively high BU incidence. Within a case/matched control study design, environmental factors associated with BU infection and spatial behaviors are investigated to uncover possible links between individual daily activity spaces and terrains of risk across disturbed landscapes. This research relies upon archival and field-collected data and analyses conducted with geographical information systems (GIS).
Contributing Partner: UNT Libraries
The Geography of Maternal Mortality in Nigeria

The Geography of Maternal Mortality in Nigeria

Date: May 2012
Creator: Ebeniro, Jane
Description: Maternal mortality is the leading cause of death among women in Nigeria, especially women aged between 15 and 19 years. This research examines the geography of maternal mortality in Nigeria and the role of cultural and religious practices, socio-economic inequalities, urbanization, access to pre and postnatal care in explaining the spatial pattern. State-level data on maternal mortality rates and predictor variables are presented. Access to healthcare, place of residence and religion explains over 74 percent of the spatial pattern of maternal mortality in Nigeria, especially in the predominantly Muslim region of northern Nigeria where poverty, early marriage and childbirth are at its highest, making them a more vulnerable population. Targeting vulnerable populations in policy-making procedures may be an important strategy for reducing maternal mortality, which would also be more successful if other socio-economic issues such as poverty, religious and health care issues are promptly addressed as well.
Contributing Partner: UNT Libraries
Spatio-temporal Variation of Nitrate Levels in Groundwater in Texas, 1970 to 2010

Spatio-temporal Variation of Nitrate Levels in Groundwater in Texas, 1970 to 2010

Date: December 2012
Creator: Rice, Susan C.
Description: This study looks at spatial variation of groundwater nitrate in Texas and its fluctuations at 10 year increments using data from the Texas Water Development Board. While groundwater nitrate increased in the Ogallala and Seymour aquifers across the time period, the overall rate in Texas appears to be declining as time progresses. However, the available data is limited. Findings show that a much more targeted, knowledge based strategy for sampling would not only reduce the cost of water quality analysis but also reduce the risk of error in these analyses by providing a more realistic picture of the spatial variation of problem contaminants, thereby giving decision-makers a clearer picture on how best to handle the reduction and elimination of problem contaminants.
Contributing Partner: UNT Libraries
Automated Treetop Detection and Tree Crown Identification Using Discrete-return Lidar Data

Automated Treetop Detection and Tree Crown Identification Using Discrete-return Lidar Data

Date: May 2013
Creator: Liu, Haijian
Description: Accurate estimates of tree and forest biomass are essential for a wide range of applications. Automated treetop detection and tree crown discrimination using LiDAR data can greatly facilitate forest biomass estimation. Previous work has focused on homogenous or single-species forests, while few studies have focused on mixed forests. In this study, a new method for treetop detection is proposed in which the treetop is the cluster center of selected points rather than the highest point. Based on treetop detection, tree crowns are discriminated through comparison of three-dimensional shape signatures. The methods are first tested using simulated LiDAR point clouds for trees, and then applied to real LiDAR data from the Soquel Demonstration State Forest, California, USA. Results from both simulated and real LiDAR data show that the proposed method has great potential for effective detection of treetops and discrimination of tree crowns.
Contributing Partner: UNT Libraries
FIRST PREV 1 2 NEXT LAST