Multi-sensor data collection and fusion using autoencoders in condition evaluation of concrete bridge decks

Abstract This paper presents a multi-sensor data collection and fusion procedure for nondestructive evaluation/testing (NDE/NDT) of concrete bridge deck. Three NDE technologies, vertical electrical impedance (VEI), ground-penetrating radar (GPR), high-definition imaging (HDI) surface crack detection, were deployed on the A neural network autoencoder was trained to quantify relationship between VEI GPR results using collected at common positions. then used increase reliability spatial resolution measurements generate data-fused condition map that showed novel characteristics. Threshold values tests obtained determine color scale in fused map. Surface cracks identified from HDI show reasonable agreement with deterioration areas Chloride concentration sound deteriorated deck consistent results.