Method for automatically generating a three-dimensional reference model as terrain information for an imaging device

1. A method for automatically generating a three-dimensional reference model as terrain information for a seeker head of an unmanned missile with a predetermined field of view comprising the steps:

a) providing a three-dimensional terrain model formed from model elements, which has been obtained using satellite and/or aerial reconnaissance;

b) specifying position data of the imaging device at least at one planned position and a direction vector from the planned position of the imaging device to a predetermined target point in the three-dimensional terrain model;

c) generating a three-dimensional reference model of the three dimensional terrain model, wherein only model elements and sections of model elements from the terrain model are incorporated into the reference model that, in a viewing direction of the direction vector from the planned position of the imaging device, are not covered by other model elements or are not located outside a field of view of the imaging device, wherein

accuracy information reflecting position accuracy of a respective model element is respectively assigned to the model elements of the three-dimensional terrain model, and

model elements with a better accuracy are used in the generation of the three-dimensional reference model, and model elements with poorer accuracy are not incorporated into the reference model.

2. The method according to claim 1 , wherein the model elements provided for incorporation into the reference model are examined for whether, in the viewing direction of the direction vector from the planned position of the imaging device, they form structures that are indistinct.

3. The method according to claim 2 , wherein

a neighborhood analysis for a respective model element is performed before the selection of a model element for the generation of the reference model, in which neighborhood analysis it is verified whether, for the respective model element, similar model elements are present in a predetermined environment to the respective model element, and

then, if a similar model element has been detected in the predetermined environment in the neighborhood analysis, the respective image element is not incorporated into the reference model or incorporated only with low accuracy information.

4. The method according to claim 3 , wherein groups of similar model elements ascertained in the neighborhood analysis are incorporated into the reference model either completely or not at all.

5. The method according to claim 1 , wherein

to verify accuracy of a detected reference model, initially at least one group of model elements is selected, and it is then determined for this group the accuracy with which the position of the imaging device can be determined by pattern comparison of this group of model elements with model elements extracted from an image captured by the imaging device, and

only model elements of those groups of model elements are incorporated into the reference model in which the aforementioned position determination is carried out with an accuracy that lies above a predetermined threshold value.

6. The method according to claim 1 , wherein the model elements have lines, intersection points of lines and areas delimited by lines.

7. The method according to claim 6 , wherein the lines reflect contour edges of objects present in an actual scene depicted by the three-dimensional terrain model.