Land subsidence monitoring using airborne LiDAR data and relevant auxiliary data

Abstract

Precise leveling has high accuracy for land subsidence monitoring, but it is time-consuming and only profile data can be captured. Airborne LiDAR (Light Detection And Ranging) acquires dense and accurate point clouds on land surface rapidly. This technique can offer the potential for comprehensive analysis of land subsidence. However, the point clouds could be located on unstable terrain objects, and the accuracy of single point in point clouds may not be suitable for the evaluation of land subsidence in regular flight scanning design. After specific flight scanning design, the extraction of point clouds on stable terrain objects, e.g. road surfaces or building rooftops, and accurate elevation determination on the representative location for land subsidence monitoring are necessary. In this study, the test area with 2 km by 10 km will be divided into 5m by 5m sub-grids and each centric elevation of the grid will be calculated by the fitting planar parameters, which is determined by least squares fitting according to the point clouds in each grid. Subsequently, three approaches for acquiring stable monitoring planes will be presented in this study: (1) to classify the point clouds on stable terrain objects by its intensity, (2) to extract the point clouds on road surfaces from the road information in the topographic maps with 1/5,000 scale, (3) to extract the point clouds on rooftops from the building information in the topographic maps with 1/5,000 scale. Finally, the land subsidence between two different periods of airborne LiDAR data from the approaches mentioned above will be evaluated and analyzed based on the collected ground truth. Copyright © (2009) by the Asian Association on Remote Sensing.