US 9,317,968 B2 · System and method for multiple hypotheses testing for surface orientation during 3D point cloud extraction from 2D imagery

Abstract

The system and methods described herein operate on a plurality of images that include multiple views of the same scene, typically from slightly different viewing angles and/or lighting conditions. One of the images is selected as a reference image. For each image ray in a non-reference image, the system and methods resample a local region from the non-reference image's space to the reference image's space. The resampling is performed multiple times, each time with a different surface orientation hypothesis. The system and methods run cross-correlation style correlators on the resampled images, evaluate correlation scores for each of the resampled images, and select the surface orientation hypothesis associated with the highest correlation score. The system and methods project a peak of the correlation surface back through a geometry model for the selected surface orientation hypothesis to determine a three-dimensional (ground) location for the image ray.

Claims14

Dependent claims

A system for extracting a plurality of three-dimensional points that represent a terrain, the three-dimensional points being extracted from first and second images, the first and second images being two-dimensional representations of the terrain generated from different viewing angles, the system comprising processing circuitry arranged to, for each ray in the first image:

form a plurality of different hypotheses for a true orientation of the terrain;
resample a local region from the first image into a space of the second image multiple times, each of the resamplings using a different one of the hypotheses from the plurality of hypotheses;
run cross-correlation style correlators on the plurality of hypotheses to produce respective pluralities of correlation scores and correlation surfaces;
select a hypothesis from the plurality of hypotheses that has the highest correlation score;
select a correlation surface corresponding with the selected hypothesis;
project a peak of the selected correlation surface through a geometry model for the selected hypothesis;
and
determine a three-dimensional point representing a location of the ray from the projected peak.

5 more elements

A system for extracting a plurality of best-guess ground locations of tie-points that represent a terrain, the tie-points being extracted from a reference image and a non-reference image, the reference and non-reference images being two-dimensional representations of the terrain generated from different viewing angles, the system comprising processing circuitry arranged to, for each tie-point:

for each of a plurality of different hypotheses, determine and store a trial ground location;
for each of the plurality of different hypotheses, determine and store a correlation score;
compare the correlation scores of the plurality of different hypotheses;
select a trial ground location, of the stored trial ground locations, that corresponds to the highest correlation score;
select a hypothesis, of the plurality of different hypotheses, that corresponds to the highest correlation score;
and
for the selected hypothesis, project a peak of a respective correlation surface through a respective geometry model to determine the best-guess ground location of the tie-point.

4 more elements

A non-transitory computer-readable storage medium storing a program for causing a computer to implement a method for extracting a plurality of ground locations of tie-points that represent a terrain, the tie-points being extracted from a reference image and a non-reference image, the reference and non-reference images being two-dimensional representations of the terrain generated from different viewing angles, the method comprising, for each tie-point:

chipping a template window from the reference image;
for each of a plurality of different hypotheses, determining and storing a guessed ground location;
for each of the plurality of different hypotheses, determining and storing a correlation score;
comparing the correlation scores of the plurality of different hypotheses;
selecting a guessed ground location, of the stored guessed ground locations, that corresponds to the highest correlation score;
selecting a hypothesis, of the plurality of different hypotheses, that corresponds to the highest correlation score;
and
for the selected hypothesis, projecting a peak of a respective correlation surface through a respective geometry model to determine the ground location of the tie-point.

5 more elements

Dependencies are inferred from the claim text.

System and method for multiple hypotheses testing for surface orientation during 3D point cloud extraction from 2D imagery

Life of this patent

System and method for multiple hypotheses testing for surface orientation during 3D point cloud extraction from 2D imagery

Life of this patent

Abstract

Claims14