IP Library › Granted Patent US 12682566
Granted Patent B2
US 12682566 · App. 18/402,893 · Granted Jul 14, 2026

Enhancing elevation models with landcover feature data

Inventors: Yuanming Shu (Toronto, CA); Shuo Tan (Campbell, CA); Dawei Zai (Toronto, CA); Ruijie Deng (Toronto, CA); Zihao Chen (Toronto, CA); Andi Dai (Toronto, CA)
Assignee: ECOPIA TECH CORPORATION
G06T17/05G06T19/20G06V10/25G06V10/40G06T2219/2004
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12682566
App. No.
18/402,893
Granted
Jul 14, 2026
Kind
B2
Abstract

Methods and systems for enhancing elevation models with landcover feature data are provided. An example method involves accessing imagery covering an area of interest that contains an above-ground feature, generating, a digital surface model of the area of interest, extracting three-dimensional landcover feature data representing the above-ground feature that contains one or more actual ground-level points at the base of the above-ground feature, and generating a digital terrain model from the digital surface model at least in part by removing the above-ground feature from the digital surface model and filling at least some of the removed area with one or more of the actual ground-level points from the three-dimensional landcover feature data.

Claims (48)

1 . A method comprising:

accessing geospatial imagery covering an area of interest, wherein the area of interest includes one or more building features and one or more above-ground natural landcover features;

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 one or more building features and the one or more above-ground natural landcover features;

extracting, from the geospatial imagery, first two-dimensional vector data that represents one or more boundaries of the one or more building features, and second two-dimensional vector data that represents one or more boundaries of the one or more above-ground natural landcover features;

generating a preliminary digital terrain model over the area of interest, wherein generating the preliminary digital terrain model involves removing from the digital surface model portions of the digital surface model that represent above-ground points bound by the first and second two-dimensional vector data;

generating first three-dimensional vector data that represents the one or more building features, wherein the first three-dimensional vector data includes one or more ground-level points at one or more bases of the one or more building features;

generating a refined digital terrain model over the area of interest, wherein generating the refined digital terrain model involves inserting the one or more ground-level points of the first three-dimensional vector data into the preliminary digital terrain model and filling the preliminary digital terrain model between the one or more ground-level points with one or more estimated ground-level points;

generating second three-dimensional vector data that represents the one or more above-ground natural landcover features based on the second two-dimensional vector data and the above-ground points bound by the second two-dimensional vector data; and

combining the refined digital terrain model with the first three-dimensional vector data and the second three-dimensional vector data to produce a structural digital surface model.

2 . The method of claim 1 , wherein:

generating the second three-dimensional vector data involves deriving a height attribute for an above-ground natural landcover feature from one or more raw data points of the digital surface model located within the area bound by the second two-dimensional vector data.

3 . The method of claim 1 , wherein:

filling the preliminary digital terrain model between the one or more ground-level points with one or more estimated ground-level points involves using a non-linear optimization process in which the one or more ground-level points of the first three-dimensional vector data are used a seed in the non-linear optimization process.

4 . The method of claim 1 , wherein generating the first three-dimensional vector data comprises:

applying a machine learning model trained to extract three-dimensional building features from the accessed geospatial imagery.

5 . The method of claim 1 , wherein the three-dimensional landcover feature comprises forest canopy.

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 one or more building features and one or more above-ground natural landcover features;

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 one or more building features and the one or more above-ground natural landcover features;

extract, from the geospatial imagery, first two-dimensional vector data that represents one or more boundaries of the one or more building features, and second two-dimensional vector data that represents one or more boundaries of the one or more above-ground natural landcover features;

generate a preliminary digital terrain model over the area of interest, wherein generating the preliminary digital terrain model involves removing from the digital surface model portions of the digital surface model that represent above-ground points bound by the first and second two-dimensional vector data;

generate first three-dimensional vector data that represents the one or more building features, wherein the first three-dimensional vector data includes one or more ground-level points at one or more bases of the one or more building features;

generate a refined digital terrain model over the area of interest, wherein generating the refined digital terrain model involves inserting the one or more ground-level points of the first three-dimensional vector data into the preliminary digital terrain model and filling the preliminary digital terrain model between the one or more ground-level points with one or more estimated ground-level points;

generate second three-dimensional vector data that represents the one or more above-ground natural landcover features based on the second two-dimensional vector data and the above-ground points bound by the second two-dimensional vector data; and

combine the refined digital terrain model with the first three-dimensional vector data and the second three-dimensional vector data to produce a structural digital surface model.

7 . The system of claim 6 , wherein:

generating the second three-dimensional vector data involves deriving a height attribute for an above-ground natural landcover feature from one or more raw data points of the digital surface model located within the area bound by the second two-dimensional vector data.

8 . The system of claim 6 , wherein:

filling the preliminary digital terrain model between the one or more ground-level points with one or more estimated ground-level points involves using a non-linear optimization process in which the one or more ground-level points of the first three-dimensional vector data are used a seed in the non-linear optimization process.

9 . The system of claim 6 , wherein generating the first three-dimensional vector data comprises:

applying a machine learning model trained to extract three-dimensional building features from the accessed geospatial imagery.

10 . The system of claim 6 , wherein the three-dimensional landcover feature comprises forest canopy.

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 one or more building features and one or more above-ground natural landcover features;

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 one or more building features and the one or more above-ground natural landcover features;

extract, from the geospatial imagery, first two-dimensional vector data that represents one or more boundaries of the one or more building features, and second two-dimensional vector data that represents one or more boundaries of the one or more above-ground natural landcover features;

generate a preliminary digital terrain model over the area of interest, wherein generating the preliminary digital terrain model involves removing from the digital surface model portions of the digital surface model that represent above-ground points bound by the first and second two-dimensional vector data;

generate first three-dimensional vector data that represents the one or more building features, wherein the first three-dimensional vector data includes one or more ground-level points at one or more bases of the one or more building features;

generate a refined digital terrain model over the area of interest, wherein generating the refined digital terrain model involves inserting the one or more ground-level points of the first three-dimensional vector data into the preliminary digital terrain model and filling the preliminary digital terrain model between the one or more ground-level points with one or more estimated ground-level points;

generate second three-dimensional vector data that represents the one or more above-ground natural landcover features based on the second two-dimensional vector data and the above-ground points bound by the second two-dimensional vector data; and

combine the refined digital terrain model with the first three-dimensional vector data and the second three-dimensional vector data to produce a structural digital surface model.

12 . The at least one non-transitory machine-readable storage medium of claim 11 , wherein:

generating the second three-dimensional vector data involves deriving a height attribute for an above-ground natural landcover feature from one or more raw data points of the digital surface model located within the area bound by the second two-dimensional vector data.

13 . The at least one non-transitory machine-readable storage medium of claim 11 , wherein:

filling the preliminary digital terrain model between the one or more ground-level points with one or more estimated ground-level points involves using a non-linear optimization process in which the one or more ground-level points of the first three-dimensional vector data are used a seed in the non-linear optimization process.

14 . The at least one non-transitory machine-readable storage medium of claim 11 , wherein generating the first three-dimensional vector data comprises:

applying a machine learning model trained to extract three-dimensional building features from the accessed geospatial imagery.

15 . The at least one non-transitory machine-readable storage medium of claim 11 , wherein the three-dimensional landcover feature comprises forest canopy.