Method and system for geo-referencing at least one sensor image

1. Method ( 700 ) for geo-referencing at least one sensor image, said method comprising the steps of:

generating ( 701 ) said at least one sensor image of a first scene with at least one sensor;

accessing ( 702 ) a 3D model of the environment comprising geo-coded 3D coordinate data and related to at least one second scene, said second scene encompassing said first scene;

matching ( 703 ) the sensor image with the 3D model to find a section of the 3D model where there is a match between the first and the second scenes;

geo-referencing ( 704 ) the sensor image based on the geo-coded 3D coordinate data of the found section of the 3D model; and

determining ( 705 ) a measure related to an uncertainty in the matching between the sensor image and the 3D model,

wherein:

the 3D model of the environment comprises a textured mesh describing the environment, the textured mesh comprising nodes interconnected via edges and surfaces boarded by the edges; and

each node and/or edge and/or surface is associated to geo-coded 3D coordinate data and texture data.

2. Method according to claim 1 , wherein the matching ( 703 ) involves positioning and rotating the sensor image in the 3D model and changing intrinsic parameters such as at least one of field of view or radial distortion.

3. Method according to claim 1 , wherein the determination ( 705 ) of the measure related to an uncertainty comprises:

comparing the sensor image with the corresponding section in the 3D model; and

determining a probability that the matching between the model and the sensor image is correct.

4. Method according to claim 1 , further comprising a step of extracting ( 706 ) coordinate information and the associated measure related to the uncertainty related to a selected point in the sensor image.

5. Method according to claim 1 , wherein the determining of the measure related to the uncertainty is based on a model uncertainty.

6. Method according to claim 1 , wherein the geo-coded 3D coordinate data is given in a geo-referenced coordinate system.

7. Method ( 700 ) for geo-referencing at least one sensor image, said method comprising the steps of:

generating ( 701 ) said at least one sensor image of a first scene with at least one sensor;

accessing ( 702 ) a 3D model of the environment comprising a textured mesh describing the environment, said 3D model comprising geo-coded 3D coordinate data and texture data, and being related to at least one second scene, said second scene encompassing said first scene;

matching ( 703 ) the sensor image with the 3D model to find a section of the 3D model where there is a match between the first and the second scenes;

geo-referencing ( 704 ) the sensor image based on the geo-coded 3D coordinate data of the found section of the 3D model; and

determining ( 705 ) a measure related to an uncertainty in the matching between the sensor image and the 3D model,

wherein the measure related to the uncertainty in the matching is determined ( 705 ) for the sensor image as a whole.

8. Method according to claim 7 , wherein the measure related to the uncertainty in the matching is determined as a distance value.

9. Method according to claim 7 , wherein the measure related to the uncertainty in the matching is determined as a percentage.

10. Method according to claim 7 , further comprising the steps of:

determining a similarity between the sensor image and the found section of the 3D model;

determining similarities between the sensor image and sections of the model adjacent to the found section; and

determining ( 705 ) the measure related to the uncertainty based on a comparison between the determined similarities.

11. Method according to claim 7 , wherein the determining ( 705 ) of a measure related to an uncertainty comprises:

comparing all points in the sensor image with a corresponding point in the model; and

forming an accuracy measure based on the distance between each respective point in the sensor image and the corresponding point in the model.

12. Method ( 700 ) for geo-referencing at least one sensor image, said method comprising the steps of:

generating ( 701 ) said at least one sensor image of a first scene with at least one sensor;

accessing ( 702 ) a 3D model of the environment comprising a textured mesh describing the environment, said 3D model comprising geo-coded 3D coordinate data and texture data, and being related to at least one second scene, said second scene encompassing said first scene;

matching ( 703 ) the sensor image with the 3D model find a section of the 3D model where there is a match between the first and the second scenes;

geo-referencing ( 704 ) the sensor image based on the geo-coded 3D coordinate data of the found section of the 3D model; and

determining ( 705 ) a measure related to an uncertainty in the matching between the sensor image and the 3D model,

wherein the measure related to an uncertainty is provided for each point in the sensor image.

13. Method according to claim 12 , wherein the determination ( 705 ) of the measure related to the uncertainty comprises:

comparing each point in the sensor image with a corresponding point in the 3D model; and

forming the measure related to the uncertainty based on the distance between each point in the sensor image and the corresponding point in the model.

14. Method ( 700 ) for geo-referencing at least one sensor image, said method comprising the steps of:

generating ( 701 ) said at least one sensor image of a first scene with at least one sensor;

accessing ( 702 ) a 3D model of the environment comprising geo-coded 3D coordinate data and related to at least one second scene, said second scene encompassing said first scene;

matching ( 703 ) the sensor image with the 3D model find a section of the 3D model where there is a match between the first and the second scenes;

geo-referencing ( 704 ) the sensor image based on the geo-coded 3D coordinate data of the found section of the 3D model; and

determining ( 705 ) a measure related to an uncertainty in the matching between the sensor image and the 3D model,

wherein the 3D model of the environment comprises a mesh describing the environment and comprising nodes interconnected by means of edges and surfaces boarded by the edges, wherein each node and/or edge and/or surface is associated to geo-coded 3D coordinate data and the associated measure related to the uncertainty.

15. Method according to claim 14 , wherein the 3D model is textured.

16. A non-transitory computer program product for geo-referencing at least one sensor image, the computer program product comprising at least one computer-readable storage medium having computer-readable program code portions embodied therein, the computer-readable program code portions comprising:

an executable portion configured for generating said at least one sensor image of a first scene with at least one sensor;

an executable portion configured for accessing a 3D model of the environment comprising geo-coded 3D coordinate data and related to at least one second scene, said second scene encompassing said first scene, said 3D model of the environment comprising a textured mesh describing the environment, the textured mesh comprising nodes interconnected via edges and surfaces boarded by the edges, each node and/or edge and/or surface being associated to geo-coded 3D coordinate data and texture data;

an executable portion configured for matching the sensor image with the 3D model find a section of the 3D model where there is a match between the first and the second scenes;

an executable portion configured for geo-referencing the sensor image based on the geo-coded 3D coordinate data of the found section of the 3D model; and

an executable portion configured for determining a measure related to an uncertainty in the matching between the sensor image and the 3D model.

17. System ( 200 ) for geo-referencing at least one sensor image, said system comprising:

at least one sensor ( 203 ) arranged to capture at least one image of the first scene;

a memory ( 204 ) for storing a 3D model of the environment comprising geo-coded 3D coordinate data, said 3D model of the environment comprising a textured mesh describing the environment, the textured mesh comprising nodes interconnected via edges and surfaces boarded by the edges, each node and/or edge and/or surface being associated to geo-coded 3D coordinate data and texture data; and

a processing unit ( 205 ) configured to access the 3D model, to match the sensor image with the 3D model so as to find a section of the 3D model where there is a match, to geo-reference the sensor image based on the geo-coded 3D coordinate data of the found section of the 3D model, and to determine a measure related to an uncertainty in the matching between the sensor image and the 3D model.