Abstract
Hidden-landslides that appeared in an earthquake are not always visible in aerial photographs. All the more because these hidden landslides are at risk of reactivation in heavy rains and snow melting times that follow the earthquake, it is often an urgent need for us to detect their presence for rational rehabilitations Recent development of remote sensing technologies allows us to observe changes in terrains in a very precise manner. However, to deal with land concervation issues, we need to extract Lagrandian soil displacements first, and then separate the hidden landslides. For the separate extraction of hidden landslides, moving average method is used herein to create a series of averages of different subsets of the full data set of the detected Lagrangian displacements, and the most appropriate square-window size for the hidden landslides is examined at Kizawa area where a 300m-long road tunnel was deformed by a hidden landslide.
