Granted Patent B2

US 12,394,029 · App. 18/082,078 · Granted Aug 19, 2025

Inventors: Dmitry Aleksandrovich Sukhov (Podolsk, RU); Nikolay Alekseevich Fedorov (Kotelniki, RU); Roman Evgenevich Sokolov (Naberezhnyye Chelny, RU)

Assignee: Y.E. Hub Armenia LLC

G06T5/80G06T5/50G06T7/344G06T2207/10032G06T2207/20081G06T2207/30184

Abstract

Claims (63)

1. A method comprising:

receiving a satellite image comprising one or more roads and a digital elevation model (DEM) comprising an elevation of each pixel of the satellite image;

receiving recorded global positioning system (GPS) data corresponding to the one or more roads;

determining a pixel in the satellite image corresponding to the one or more roads;

determining whether a location of the pixel in the satellite image corresponds to the GPS data;

after determining that the location of the pixel in the satellite image does not correspond to the GPS data, determining a set of pixels of the DEM that surround the pixel in the satellite image;

randomly selecting elevations for each pixel of the set of pixels;

applying the randomly selected elevations to the DEM;

after applying the randomly selected elevations to the DEM, determining an adjusted location of the pixel;

determining whether the adjusted location of the pixel in the satellite image corresponds to the one or more roads; and

after determining that the adjusted location of the pixel in the satellite image corresponds to the one or more roads, storing the randomly selected elevations in the DEM.

2. The method of claim 1 , further comprising:

determining a second set of pixels corresponding to roads in the satellite image; and

selecting, from the second set of pixels corresponding to roads in the satellite image, the pixel in the satellite image.

3. The method of claim 2 , wherein determining the second set of pixels corresponding to roads in the satellite image comprises:

inputting the satellite image into a trained machine learning algorithm (MLA); and

outputting, by the MLA, the second set of pixels corresponding to roads in the satellite image.

4. The method of claim 3 , wherein the trained MLA was trained using a set of satellite images, wherein each satellite image in the set of satellite images comprises a label indicating which pixels of the respective satellite image correspond to roads in the respective satellite image.

5. The method of claim 1 , further comprising determining a rational polynomial coefficient (RPC) corresponding to the pixel in the satellite image.

6. The method of claim 5 , wherein determining whether the location of the pixel in the satellite image corresponds to the GPS data comprises:

determining, based on the RPC, a GPS coordinate corresponding to the pixel in the satellite image; and

comparing the GPS coordinate of the pixel in the satellite image to the GPS data to determine whether the GPS coordinate corresponds to a road.

7. The method of claim 6 , wherein determining whether the location of the adjusted pixel in the satellite image corresponds to the one or more roads comprises:

determining a second RPC corresponding to the adjusted pixel in the satellite image;

determining, based on the second RPC, a second GPS coordinate corresponding to the adjusted pixel in the satellite image; and

comparing the second GPS coordinate of the adjusted pixel in the satellite image to the GPS data to determine whether the second GPS coordinate corresponds to a road.

8. The method of claim 1 , wherein the recorded GPS data comprises GPS coordinates recorded by a plurality of electronic devices traveling on the one or more roads.

9. The method of claim 8 , wherein the plurality of electronic devices are associated with vehicles traveling on the one or more roads.

10. The method of claim 1 , wherein the recorded GPS data is generated by:

receiving GPS coordinates recorded by a plurality of vehicles traveling on the one or more roads; and

removing outliers from the GPS coordinates.

11. A method comprising:

receiving a satellite image comprising one or more roads and a digital elevation model (DEM) comprising an elevation of each pixel of the satellite image;

receiving recorded global positioning system (GPS) data corresponding to the one or more roads;

determining, based on the GPS data, a shift to apply to the satellite image;

selecting a pixel in the DEM;

determining, based on pixels surrounding the pixel, a plurality of candidate pixel values of the pixel;

determining, for each candidate pixel value of the plurality of candidate pixel values, an amount of error corresponding to the respective candidate pixel value;

selecting a candidate pixel value having a lowest amount of error of the plurality of candidate pixel values; and

generating, based on the selected candidate pixel value, an adjusted DEM.

12. The method of claim 11 , further comprising, generating, based on the adjusted DEM, a corrected satellite image.

13. The method of claim 11 , further comprising receiving rational polynomial coefficients (RPCs) corresponding to the satellite image, and wherein determining the shift to apply to the satellite image comprises:

generating a plurality of variants of the RPCs;

determining, for each variant of the plurality of variants, an amount of error based on the GPS data;

selecting a variant of the plurality of variants having a smallest amount of error, wherein the variant corresponds to the shift; and

applying the shift to the RPCs.

14. The method of claim 11 wherein determining the plurality of candidate pixel values comprises applying a smoothing algorithm to the pixel and the pixels surrounding the pixel.

15. The method of claim 11 , wherein determining the amount of error comprises determining an indication of overlap between the one or more roads and the GPS data.

16. A method comprising:

receiving a satellite image comprising one or more roads, a digital elevation model (DEM) comprising an elevation of each pixel of the satellite image;

receiving recorded global positioning system (GPS) data corresponding to the one or more roads;

selecting a pixel in the DEM;

generating a plurality of candidate pixel values of the pixel;

determining, for each candidate pixel value of the plurality of candidate pixel values, an amount of error corresponding to the respective candidate pixel value;

selecting a candidate pixel value having a lowest amount of error of the plurality of candidate pixel values; and

generating, based on the selected candidate pixel value, an adjusted DEM.

17. The method of claim 16 , wherein determining the amount of error for a respective candidate pixel value comprises:

determining a set of pixels surrounding the pixel;

determining, for the set of pixels, an amount of overlap between the GPS data and the set of pixels; and

determining, based on the amount of overlap, the amount of error for the respective candidate pixel value.

18. The method of claim 16 , wherein each candidate pixel value is within a pre-determined range of the pixel.

19. The method of claim 16 , further comprising, generating, based on the adjusted DEM, a corrected satellite image.

20. The method of claim 11 , wherein generating the adjusted DEM comprises replacing the pixel with the selected candidate pixel value.

Assignments (5)

Priority Claims (1)

Continuity (1)

References Cited (12)