The non-ballasted technology has yielded numerous new railway tracks, especially in tunnels. The "low vibration track" (LVT) is a system developed in 1980s, and it today covers more than 1100km of track around the world, including the Channel tunnel. This system contains a concrete block based on a micro-cellular pad all integrated in a rubber boot embedded in the slab of unreinforced concrete. It is difficult to determine the block degradation because it is hidden by the surrounding. This article presents a fast and efficient detection method for the block monitoring. Basically, the block damage occurs in two types: vertical and horizontal cracks. Based on the modal analysis, this study firstly presents the link between the eigenmodes and the block damages by measurements and modeling. The 3D modeling shows that the first mode (flexion) and the second one (torsion) are influenced by two types of damage. Then, we present the database of 284 damage blocks to analyze the relation between eigenmode and damage. Finally, the detection method has been developed to identify block damage level by using modal analysis of vibration response. A robot called COBRA has been developed for the Channel tunnel which automatizes the block monitoring at a high speed.