Accuracy assessment of digital elevation models generated using the Erdas Imagine OrthoMAX digital photogrammetric system

The users of many digital photogrammetric systems are allowed a degree of control over the automated digital elevation model (DEM) generation software with a set of strategy parameters. These parameters control the search and quality control characteristics of the algorithm; hence the wrong choice of parameters can have a significant detrimental effect on the accuracy of the DEM. This paper describes the strategy parameters used in the Erdas Imagine OrthoMAX digital photogrammetric system and a set of tests used to define the optimum set of parameters for a project. The algorithm employs an area correlator and a hierarchical approach in the matching process. The results show that a 35 per cent increase in accuracy can be achieved through the process of parameter optimization. They also demonstrate that changes to the strategy parameters have a similar effect on the results for all areas of a set of imagery used in the test, which is of particular note because the imagery contains a diverse range of landcover types. The paper additionally provides a description of a system currently being developed which will automatically prescribe the correct set of parameters for a project.