Inspection system for the conservation of flexible pavements through the Cascade Trainer algorithm and the use of drone (#1208)

Authors

Poma, Marvin

Marin, Alexis

Silvera, Manuel

Palacios-Alonso, Daniel

Abstract

The pavement inspection process is of utmost importance to know its condition and thus make decisions regarding its conservation and maintenance. This research proposes a methodology to optimize the pavement inspection process based on the ASTM D6433 standard through the use of drones and the Cascade Trainer algorithm. Images of pavement pathologies were used to train the algorithm and with it they will carry out the pavement inspection in the videos captured by the DJI mini 4 pro drone. The algorithm demonstrated high performance with an average accuracy of 96%, a cost reduction of 77% compared to the “American Society for Testing and Materials” Method (ASTM D6433) and a performance capable of measuring 600 more meters per workday. Finally, the results of this research can be useful to promote a proposal for improvement in the inspection of flexible pavements for their conservation through automated and high-precision methods