Master of Science in Civil Engineering (MSCE)
Civil and Environmental Engineering
The main focus of this study is to improve the data analysis tools used in performance monitoring and level of service assessment of freeway systems. The proposed study presents a methodology to develop new second-order statistical measures that are derived from texture characterization techniques in the field of digital image analysis. The new measures are capable of extracting properties such as smoothness, homogeneity, regularity, and randomness in traffic behavior from the spatio-temporal traffic contour maps. To study the new performance measures a total of 14270, 15-min traffic contour maps were generated for a section of 3.4 miles of I-4 in Orlando, Florida for 24 hours over a period of 5 weekdays. A correlation matrix was examined using the obtained measures for all the constructed maps, which is used to check for information redundancy. This resulted in retaining a set of three second-order statistical measures: angular second moment (ASM), contrast (CON), and entropy (ENT). The retained measures were analyzed to examine their sensitivity to various traffic conditions, expressed by the overall mean speed of each contour map. The measures were also used to evaluate level of service for each contour map. The sensitivity analysis and level of service criteria can be implemented in real time using a stand-alone module that was developed in this study. The study also presents a methodology to compare the traffic characteristics of various congested conditions. To examine the congestion characteristics, a total of 10,290 traffic contour maps were generated from a 7.5-mile section of the freeway for a period of 5 weekdays.
Document Availability at the Time of Submission
Release the entire work immediately for access worldwide.
Kotha, Prashanth, "A methodology for deriving performance measures from spatio-temporal traffic contour maps using digital image analysis procedures" (2003). LSU Master's Theses. 693.