SCANNING TUNNELS WITH TWO VERY HIGH-RESOLUTION LASER DEVICES AND A STACKER
In this paper, we analyse the potentialities of the LCMS (Laser Crack Measurement System) sensors for tunnel inspection. Two of these sensors have been integrated into a flexible prototype developed by Cerema, allowing data from the entire tunnel to be recorded dynamically in a few passes. Thanks to this system, we were able to acquire, in a limited time, range and intensity images of the complete vaults of four tunnel tubes, with high spatial resolution, during experimental campaigns. We present image processing methods to take advantage of this type of image. These approaches initially involve improving data visualization and reconstructing a local 3D model. We propose some preliminary approaches to correct the distance image in order to better highlight local variations in depth. These data are currently being deployed for learning neural networks for defect detection.