Statistical analysis of an in-vehicle image-based data collection method for assessing airport pavement condition

This study presents a comprehensive comparative statistical analysis to validate a novel in-vehicle image-based method for collecting pavement condition data in airport environments. It highlights the method's potential to address key challenges faced by airport pavement managers, such as the need for continuous maintenance and the demand for fast, effective, and reliable inspection procedures. The in-vehicle system integrates laser scanning systems, image capture, and georeferencing devices to collect pavement distress data, and its accuracy and reliability are evaluated statistically. The primary objective is to validate and enhance this novel inspection approach, which shows strong potential as an effective alternative for comprehensive pavement evaluation, enabling continuous, rapid monitoring and the analysis of trends. Validation was performed by means of a detailed statistical comparison of pavement distress density on the main runway of Amílcar Cabral International Airport, Sal Island, Cape Verde, based on data collected using the proposed in-vehicle and the traditional on-foot inspection methods. Non-parametric repeated measures analysis (nparLD) showed statistically similar results between methods for 9 of 12 distress type-severity combinations (4 types × 3 levels), especially for medium and high severity cases, and that pavement section and method-section factors were significant in 10 and 9 of 12 cases, respectively, indicating spatial variability. Kruskal-Wallis tests were applied to each method separately. Significant section-based differences were found in 11 of 12 cases for the traditional method and in 2 of 12 cases for the in-vehicle image-based method, indicating greater sensitivity of the on-foot inspection to spatial variation in distress distribution. These findings support the statistical validation of the proposed method for practical application in airport pavement management. Furthermore, the comprehensive analysis, which included correlation and autocorrelation studies, revealed a bias in severity level assignment during traditional on-foot inspections. The findings highlight time-efficiency gains with the image-based method and suggest improvements, such as enhancing image quality and providing inspector training to increase the accuracy of severity level classification. These results offer valuable insights for airport pavement managers, contributing to improved safety, operational efficiency, and resilience in the face of growing air traffic demands. • Highlights the importance of accurate and reliable pavement condition data supporting standards. • Introduces an innovative image-based data collection method for airport pavement assessment. • Employs advanced statistics to validate the proposed method against traditional inspection. • Shows that multi-functional ground vehicle method is reliable as traditional inspections. • Flags severity classification variations, highlights biases, and recommends improvements.