Evaluación de la condición de envejecimiento en pavimentos flexibles mediante técnicas de espectroradiometría y clasificación en imágenes multiespectrales.

Abstract

This research presents a methodology for network level, which allows obtaining higher performances compared to the conventional PCI (Pavement Condition Index) technique in the assessment of deterioration in flexible pavements. In the methodological development a multispectral image obtained from the sensor WorldView - 3 was used for a study area located in a section of the Main Highway 80 in Bogotá DC, the study also counted with spectral information of the different deterioration conditions of flexible pavement, this information was acquired on field using a FieldSpect 4ADS (Analytical spectral Devices, Inc.) spectroradiometer. The image classification was performed using the SAM (spectral Angle Mapper) pixel oriented spectral classification algorithm. In terms of thematic accuracy, the overall accuracy of the classification obtained in the evaluation was 85% and the value of Kappa statistic obtained was 70%, and from a practical point of view, it was observed that the methodology allows estimating the state of the pavement with a good degree of fit between the predicted and actual classes to a road network level.