 This study compares two software packages, Orthoengine and RPC Stereo Processor, for generating digital surface models, DSMs, from very high-resolution satellite stereo images over different land cover objects such as agricultural plastic greenhouses, bare soil, and urban areas. The results show that RSP produces significantly better scores in DSM completeness than PCI for all sensors and land cover objects, with a percentage improvement of approximately 2%, 18%, and 26% for bare soil, greenhouses, and urban areas respectively. However, both software packages achieved similar vertical accuracy for the different land cover objects and tested sensors. This article was authored by Manuel A. Ogila, Abderahim Nemaoui, Fernando J. Ogila, and others.