High resolution digital elevation model generation by semi global matching approach

Abstract

The photogrammetric point cloud is a revolutionary product that represents an important breakthrough in Photogrammetry. The purpose of this research is to establish the workflow in the generation of Digital Elevation Model (DEM) from the photogrammetric point cloud. The photogrammetric datasets were acquired using Digital Modular Camera 01 (DMC 01) over the Putrajaya area. The main steps for Digital Elevation Model generation using the digital photogrammetric software were involved post processing image and Global Positioning System (GPS) and inertial measurement unit (IMU), aerial triangulation, and photogrammetric point cloud generation, point cloud editing and surfacing. The whole process was successfully performed and the output was delivered in the raster DEM format. The generated Digital Elevation Model were statistically compared with DEM’s available from airborne Light, Detect and Ranging (LiDAR) technique and ground control points (GCP’s) to measure the accuracy of DEM produced by Semi Global Matching(SGM) using Root Mean Square Error (RMSE) and F-test analysis. Based on the analysis, horizontal accuracy meets the expected accuracy; 0.5-1.0 Ground Sampling Distance (GSD) of images and sub-meter accuracy were obtained for vertical accuracy. In closing, elevation data generated from the SGM approach with this accuracy meet the standard practice of Department Survey and Mapping Malaysia (DSMM) for vertical direction (height) requirement.