In this thesis an algorithm is described, which allows the generation of ortho images of cylindrical and conical surfaces. For this, the process of development is added to the ortho image calculation process, so that developable surfaces can be reproduced in the ortho images without distortions. Additionally, a digital elevation modell (DEM) is used, which enables the geometrically correct reproduction of reliefs in the ortho images. The presented algorithm was implemented in Matlab® and is applied on two objects. One of these objects is a simple pillar, which can be described by one single quadric surface. The second object is more complex and is here modelled by two circular cylinders. The first object is both, visually and numerically checked. For the numerical checks, check points are measured in the original images as well as in the resulting ortho image mosaic. Out of the measurements in the original images the 3D point coordinates are determined. These 3D point coordinates are developed with respect to the reference surface and finally the coordinates of those developed points are compared with the points which were directly measured in the mosaic. Then the differences between the coordinate values are used for the assessment of the used methods. For the first object a circular cylinder as well as a circular cone are used for modelling the object. For both variants the orthophotos are calculated with and without using a digital elevation model. The visual as well as the numerical comparision of the different variants show, that it is not important, which reference surface is chosen, as long as a digital elevation model is used. If a DEM is used, the rms-values of the differences in the check point coordinates range between 0,26 mm (1,05 pixel) and 0,30 mm (1,20 pixel) for both reference surfaces. If no DEM is used, the reference surface must model the object as good as possible. Otherwise large errors are introduced. Here the circular cone models the object best. For this reference surface the obtained rms-values of the differences in the check point coordinates are 0,98 mm (3,9 Pixel) and 1,69 mm (6,74 Pixel), dependent on the regarded coordinate direction. For the cylindrical reference surface these rms-values are larger: 2,26 mm (9,03 pixel) and 4,99 mm (19,97 pixel) are obtained, dependent on the regarded coordinate direction. By applying the algorithm on the second, more complex object, it finally can be shown, that the here implemented functions can be newly combined. By doing so, more complex objects can be handled by the algorithm. If less quality is accepted in the result, the introduced program produces a visually appealing result.