Enhancing elevation models with landcover feature data
Methods and systems for enhancing elevation models with landcover feature data are provided. An example method involves accessing geospatial imagery covering an area of interest that includes a three-dimensional landcover feature, generating a digital surface model that includes one or more above-ground points corresponding to the three-dimensional landcover feature, extracting two-dimensional vector data from the geospatial imagery that represents a two-dimensional boundary of the three-dimensional landcover feature, generating a digital terrain model by removing from the digital surface model a portion that is bound by the two-dimensional vector data, generating three-dimensional vector data representing the three-dimensional landcover feature, inserting one or more ground-level points of the three-dimensional vector data into the digital terrain model, and filling the digital terrain model where the portion of the digital surface model was removed with estimated ground-level points of the three-dimensional vector data as a seed in a non-linear optimization process.
1. A method for generating a refined digital terrain model, the method comprising:
accessing geospatial imagery covering an area of interest, wherein the area of interest includes a three-dimensional landcover feature;
generating, from the geospatial imagery, a digital surface model over the area of interest, wherein the digital surface model includes one or more above-ground points corresponding to the three-dimensional landcover feature;
extracting, from the geospatial imagery, two-dimensional vector data that represents a two-dimensional boundary of the three-dimensional landcover feature;
generating a digital terrain model over the area of interest, wherein generating the digital terrain model involves removing from the digital surface model a portion of the digital surface model that is bound by the two-dimensional vector data;
generating three-dimensional vector data that represents a three-dimensional model of the three-dimensional feature, wherein the three-dimensional vector data includes one or more ground-level points at a base of the three-dimensional landcover feature;
inserting the one or more ground-level points of the three-dimensional vector data into the digital terrain model; and
filling the digital terrain model where the portion of the digital surface model was removed with one or more estimated ground-level points, wherein filling the digital terrain model portion involves using a non-linear optimization process in which the one or more ground-level points of the three-dimensional vector data are used as a seed in the non-linear optimization process.
2. The method of claim 1 , wherein the three-dimensional landcover feature comprises a building.
3. The method of claim 1 , wherein generating the three-dimensional vector data that represents a three-dimensional model of the three-dimensional feature comprises:
applying a machine learning model to extract the three-dimensional vector data from multiview imagery, wherein the multiview imagery includes at least some of the geospatial imagery that was used to generate the two-dimensional vector data.
4. The method of claim 3 , wherein applying the machine learning model to extract the three-dimensional vector data comprises using the digital surface model as an initial guess as to a height of the three-dimensional landcover feature.
5. The method of claim 3 , wherein applying the machine learning model to extract the three-dimensional vector data comprises locating the two-dimensional vector data that represents the two-dimensional boundary of the three-dimensional landcover feature in a three-dimensional coordinate space established with respect to the multiview imagery.
6. A system comprising one or more computing devices configured to:
access geospatial imagery covering an area of interest, wherein the area of interest includes a three-dimensional landcover feature;
generate, from the geospatial imagery, a digital surface model over the area of interest, wherein the digital surface model includes one or more above-ground points corresponding to the three-dimensional landcover feature;
extract, from the geospatial imagery, two-dimensional vector data that represents a two-dimensional boundary of the three-dimensional landcover feature;
generate a digital terrain model over the area of interest, wherein generating the digital terrain model involves removing from the digital surface model a portion of the digital surface model that is bound by the two-dimensional vector data;
generate three-dimensional vector data that represents a three-dimensional model of the three-dimensional feature, wherein the three-dimensional vector data includes one or more ground-level points at a base of the three-dimensional landcover feature;
insert the one or more ground-level points of the three-dimensional vector data into the digital terrain model; and
fill the digital terrain model where the portion of the digital surface model was removed with one or more estimated ground-level points, wherein filling the digital terrain model portion involves using a non-linear optimization process in which the one or more ground-level points of the three-dimensional vector data are used as a seed in the non-linear optimization process.
7. The system of claim 6 , wherein the three-dimensional landcover feature comprises a building.
8. The system of claim 6 , wherein generating the three-dimensional vector data that represents a three-dimensional model of the three-dimensional feature comprises:
apply a machine learning model to extract the three-dimensional vector data from multiview imagery, wherein the multiview imagery includes at least some of the geospatial imagery that was used to generate the two-dimensional vector data.
9. The system of claim 8 , wherein applying the machine learning model trained to extract the three-dimensional vector data comprises using the digital surface model as an initial guess as to a height of the three-dimensional landcover feature.
10. The system of claim 8 , wherein applying the machine learning model to extract the three-dimensional vector data comprises locating the two-dimensional vector data that represents the two-dimensional boundary of the three-dimensional landcover feature in a three-dimensional coordinate space established with respect to the multiview imagery.
11. At least one non-transitory machine-readable storage medium comprising instructions that when executed cause one or more processors to:
access geospatial imagery covering an area of interest, wherein the area of interest includes a three-dimensional landcover feature;
generate, from the geospatial imagery, a digital surface model over the area of interest, wherein the digital surface model includes one or more above-ground points corresponding to the three-dimensional landcover feature;
extract, from the geospatial imagery, two-dimensional vector data that represents a two-dimensional boundary of the three-dimensional landcover feature;
generate a digital terrain model over the area of interest, wherein generating the digital terrain model involves removing from the digital surface model a portion of the digital surface model that is bound by the two-dimensional vector data;
generate three-dimensional vector data that represents a three-dimensional model of the three-dimensional feature, wherein the three-dimensional vector data includes one or more ground-level points at a base of the three-dimensional landcover feature;
insert the one or more ground-level points of the three-dimensional vector data into the digital terrain model; and
fill the digital terrain model where the portion of the digital surface model was removed with one or more estimated ground-level points, wherein filling the digital terrain model portion involves using a non-linear optimization process in which the one or more ground-level points of the three-dimensional vector data are used a seed in the non-linear optimization process.
12. The at least one non-transitory machine-readable storage medium of claim 11 , wherein the three-dimensional landcover feature comprises a building.
13. The at least one non-transitory machine-readable storage medium of claim 11 , wherein generating the three-dimensional vector data that represents a three-dimensional model of the three-dimensional feature comprises:
apply a machine learning model to extract the three-dimensional vector data from multiview imagery, wherein the multiview imagery includes at least some of the geospatial imagery that was used to generate the two-dimensional vector data.
14. The at least one non-transitory machine-readable storage medium of claim 13 , wherein applying the machine learning model trained to extract the three-dimensional vector data comprises using the digital surface model as an initial guess as to a height of the three-dimensional landcover feature.
15. The at least one non-transitory machine-readable storage medium of claim 13 , wherein applying the machine learning model trained to extract the three-dimensional vector data comprises locating the two-dimensional vector data that represents the two-dimensional boundary of the three-dimensional landcover feature in a three-dimensional coordinate space established with respect to the multiview imagery.