System for correcting RPC camera model pointing errors using 2 sets of stereo image pairs and probabilistic 3-dimensional models

摘要

A modeling engine (ME) for generating or “bootstrapping” a three dimensional edge model (3DEM) from two stereo pairs of images and correcting engine (CE) for correcting a camera model associated with an image are provided. The ME back projects edge detected images into 3DEMs using camera models associated with the stereo images, updates and stereo updates voxel probabilities in the 3DEMs with back projections of the edge detected images, and then merges the 3DEMs together to create a “sparse” 3DEM. The CE calculates a registration solution mapping an edge detected image of an image to an expected edge image. The expected edge image is a projection of a 3DEM using a camera model associated with the image. The CE corrects the camera model by applying the registration solution to the camera model based on determining whether the registration solution is a high confidence registration solution.

主权项

1. A method for generating a three dimensional edge model from two stereo pairs of images of an area of interest, the method comprising;
in a three dimensional edge modeling engine: generating a first edge detected image from a first stereo image, the first edge detected image representing edges detected in the first stereo image; back projecting the first edge detected image into a first three dimensional edge model using a camera model associated with the first stereo image, the first three dimensional edge model being a volume of voxels in which each voxel has a probability, referred to as a voxel probability, that a given back projection of the first edge detected image intersecting a subject voxel is an edge; updating the voxel probabilities in the first three dimensional edge model with back projections of the first edge detected image; generating a second edge detected image from a second stereo image, the second edge detected image representing edges detected in the second stereo image; back projecting the second edge detected image into the first three dimensional edge model using a camera model associated with the second stereo image; stereo updating the voxel probabilities in the first three dimensional edge model with back projections of the second edge detected image; for a second stereo pair, repeating the foregoing resulting in a second three dimensional edge model; and merging the first three dimensional edge model and second three dimensional edge model resulting in a merged three dimensional edge model.