There are many commercial software that can be used to perform Digital Surface Model (DSM) extraction from stereo satellite imagery based on Rational Function Model (RFM), however the license price is quite expensive. Therefore, this research aims to build the new software prototype for DSM extraction from stereo satellite imagery based on RFM, by using free development tool and integrating various components or libraries that are non-commercial (free and/or open source). The method applied in the software prototype consists of 3 main steps, i.e. conjugate points collection manually and automatically using image matching with RFM based Cross Correlation method, Forward RFM based Virtual Intersection to perform 3D reconstruction of conjugate points, and interpolate data conjugate points with Ordinary Kriging method to establish DSM in the raster format. Stereo satellite imagery of WorldView-2 which is equipped with Rational Polynomial Coefficients (RPCs) from the vendor was applied to verify the DSM accuracy that obtained from the software prototype. Based on RFM formed from the RPCs can be obtained DSM with accuracy of 2.47639575534581 meters in planimetry (CE90) and 4.21487430387638 meters in height (LE90).

DSM extraction, stereo imagery, rational function model (RFM), rational polynomial coefficients (RPCs), WordView-2, software development