Close-range photogrammetric data for landslide monitoring on slope area

Abstract

Malaysia has seen at least 40 cases of landslide due to slopes failures from 1961 to 2008 which have claimed at least 128 lives. This emphasizes the importance of monitoring such slopes for safety and maintenance purposes. However the monitoring of them could be both expensive and dangerous. Therefore this study provides alternative of monitoring the slope using a Digital Elevation Model (DEM) derived from close-range photogrammetric data. Mass movement on the slope area can be detected by comparison of DEMs derived from two or more epochs data. In this study the DEMs were generated from stereo photographs captured from two different times (13/05/2007 and 07/04/2008). This method used commercial pocket camera for its data capture, thus indicates its low cost and its suitability for monitoring inaccessible and risky areas remotely. Restitution of stereo model is performed by involving Ground Control Points (GCP) to yield camera parameters, positions and orientations of each photo. Having these parameters, positions of all sampling points on stereo photos can be transformed into ground coordinate system. Terrain data sampling points in this research were performed by regular sampling methods. To measure the quality of the generated DEM, they were compared with a set of data measured by reflector-less Total Station (TS). The result yield DEMs with centimeters accuracy in elevation and less than 0.5 % in volume difference. The result shows that for the period of one year there were significant terrain changes on ground surface. It was indicated by distribution of gain and loss areas on the slope, and extreme changes in longitudinal and cross sections profiling.