Artificial intelligence (AI) system and methods for generating estimated height maps from electro-optic imagery
An artificial intelligence (AI) system for geospatial height estimation may include a memory and a processor cooperating therewith to store a plurality of labeled predicted electro-optic (EO) image classified objects having respective elevation values associated therewith in a semantic label database, and train a model using trained EO imagery and the semantic label database. The processor may further estimate height values within new EO imagery for a geographic area based upon the trained model, and generate an estimated height map for the geographic area from the estimated height values and output the estimated height map on a display.
1. An artificial intelligence (AI) system for geospatial height estimation comprising:
a memory and a processor cooperating therewith to
store a plurality of labeled predicted electro-optic (EO) image classified objects having respective elevation values associated therewith in a semantic label database,
train a model using trained EO imagery and the semantic label database,
estimate height values within new EO imagery for a geographic area based upon the trained model and using a model selected from a plurality of models using game theory optimization (GTO), and
generate an estimated height map for the geographic area from the estimated height values and output the estimated height map on a display.
2. The AI system of claim 1 wherein the plurality of models comprises a plurality of stochastic gradient descent models (SGDMs).
3. The AI system of claim 2 wherein the GTO is based upon at least one reward matrix.
4. The AI system of claim 3 wherein the reward matrix further includes the classified objects from the semantic label database and a window of pixels surrounding a given pixel for which the height is to be estimated.
5. The AI system of claim 2 wherein at least one of the SGDMs comprises an Adaptive Moment Estimation (Adam) model.
6. The AI system of claim 2 wherein at least one of the SGDMs comprises a Root Mean Square Propagation (RMSProp) model.
7. The AI system of claim 1 wherein the processor is further configured to perform an accuracy assessment of the estimated height values based upon a non-EO height reference data set for the geographic area.
8. An artificial intelligence (AI) system for geospatial height estimation comprising:
a memory and a processor cooperating therewith to
store a plurality of labeled predicted electro-optic (EO) image classified objects having respective elevation values associated therewith in a semantic label database,
train a model using trained EO imagery and the semantic label database,
estimate height values within new EO imagery for a geographic area based upon the trained model and at least one stochastic gradient descent model (SGDM),
perform an accuracy assessment of the estimated height values based upon a non-EO height reference data set for the geographic area, and
generate an estimated height map for the geographic area from the estimated height values and output the estimated height map on a display.
9. The AI system of claim 8 wherein the at least one SGDM comprises a plurality of SGDMs; and wherein the processor is configured to perform a game theory optimization (GTO) on outputs of the plurality of SGDMs.
10. The AI system of claim 9 wherein the GTO is based upon at least one reward matrix including the classified objects from the semantic label database and a window of pixels surrounding a given pixel for which the height is to be estimated.
11. The AI system of claim 9 wherein at least some of the SGDMs comprise an Adaptive Moment Estimation (Adam) model and a Root Mean Square Propagation (RMSProp) model.
12. An artificial intelligence (AI) method for geospatial height estimation comprising:
storing a plurality of labeled predicted electro-optic (EO) image classified objects having respective elevation values associated therewith in a semantic label database;
training a model using trained EO imagery and the semantic label database;
estimating height values within new EO imagery for a geographic area based upon the trained model and using a model selected from a plurality of models using game theory optimization (GTO); and
generating an estimated height map for the geographic area from the estimated height values and output the estimated height map on a display.
13. The AI method of claim 12 wherein the plurality of models comprises a plurality of stochastic gradient descent models (SGDMs).
14. The AI method of claim 13 wherein the GTO is based upon at least one reward matrix including the classified objects from the semantic label database and a window of pixels surrounding a given pixel for which the height is to be estimated.
15. The AI method of claim 12 further comprising performing an accuracy assessment of the estimated height values based upon a non-EO height reference data set for the geographic area.
16. A non-transitory computer-readable medium having computer-executable instructions for causing a processor to perform steps comprising:
storing a plurality of labeled predicted electro-optic (EO) image classified objects having respective elevation values associated therewith in a semantic label database;
training a model using trained EO imagery and the semantic label database;
estimating height values within new EO imagery for a geographic area based upon the trained model and using a model selected from a plurality of models using game theory optimization (GTO); and
generating an estimated height map for the geographic area from the estimated height values and output the estimated height map on a display.
17. The non-transitory computer-readable medium of claim 16 wherein the plurality of models comprises a plurality of stochastic gradient descent models (SGDMs).
18. The non-transitory computer-readable medium of claim 17 wherein the GTO is based upon at least one reward matrix including the classified objects from the semantic label database and a window of pixels surrounding a given pixel for which the height is to be estimated.
19. The non-transitory computer-readable medium of claim 16 further having computer-executable instructions for causing the processor to perform an accuracy assessment of the estimated height values based upon a non-EO height reference data set for the geographic area.
20. An artificial intelligence (AI) system for geospatial height estimation comprising:
a memory and a processor cooperating therewith to
store a plurality of labeled predicted electro-optic (EO) image classified objects having respective elevation values associated therewith in a semantic label database,
train a model using trained EO imagery and the semantic label database,
estimate height values within new EO imagery for a geographic area based upon the trained model,
perform an accuracy assessment of the estimated height values based upon a non-EO height reference data set for the geographic area, and
generate an estimated height map for the geographic area from the estimated height values and output the estimated height map on a display.
21. The AI system of claim 20 wherein the processor is configured to estimate height values from the new EO imagery also based upon at least one stochastic gradient descent model (SGDM).