摘要: |
Variability of pavement surface distress data collection has always been an area of significant concern. When conducting evaluations of distress data manually (with raters observing pavements in question, interpreting what they see, and recording on paper) the process is subject to human errors. To minimize the impact of such human errors on these important pavement performance data, sophisticated equipment has been developed to eliminate as much of the human intervention as possible. Such technology is not without its own limitations of precision and bias. With both methodologies being used for the collection of surface distress data for the long-term pavement performance (LTPP) program, questions regarding precision and bias have been identified. In attempting to define the variability of the data for incorporation in stochastic analyses, it has become apparent how diverse and complex these distress data truly are. To adequately quantify the precision and bias, a detailed experiment was designed to evaluate the errors inherent in the different distress data collection methodologies. The facet of the experiment reported targets the variability of human distress surveyors and the biases associated with conducting surveys from film, using a slightly different projection system. Specifically, a collection of surveyors was assembled to establish the variability associated with experienced raters versus novice raters, engineers versus engineering technicians, and teams versus individuals. |