Steering Mosaic Cut Lines via Ground Confidence Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for creating ortho-mosaic images from oblique aerial imagery fail to maintain the natural appearance and geographic accuracy, as they attempt to align pixels to a rectilinear grid, leading to unnatural distortions and difficulties in navigation and interpretation.
Innovation Solution
An automated method for steering mosaic cut lines along preferred routes to form an output mosaic image, using a ground confidence map to determine the quality of coverage and alignment of source images, which allows for the creation of a geo-referenced image that maintains the natural appearance and accuracy of oblique imagery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If pixels are aligned to a rectilinear grid to create ortho-mosaic images, then geographic accuracy is improved, but the natural appearance of oblique imagery is distorted and lost
Solution Approach 1:
Instead of forcing oblique images to conform to a rectilinear grid (traditional approach), the patent inverts the approach by allowing the output mosaic to maintain the natural oblique perspective and geometry of the source images. The system steers cut lines through ground areas rather than warping images to fit a grid, preserving the natural appearance while still achieving geographic accuracy through proper stitching and ground confidence mapping.
Solution Approach 2:
The patent changes the fundamental parameter of how images are combined - from mathematical projection to rectilinear grid (traditional ortho-mosaic) to steered cut-line assembly that preserves original image geometry. By using ground confidence maps to guide cut-line steering through valid ground areas, the system achieves both geographic accuracy and natural appearance simultaneously.
2Productivity
If automated methods are used to create mosaics without human intervention, then productivity is improved, but the ability to handle complex cut-line routing through three-dimensional objects is reduced
Solution Approach 1:
The patent introduces ground confidence maps as an intermediary data structure that guides the automated cut-line steering process. These maps encode knowledge about ground versus non-ground areas, allowing the automated system to reliably navigate complex routing decisions without human intervention. The ground confidence map acts as a mediator between the automated stitching algorithm and the complex geometric realities of the terrain.
Solution Approach 2:
The system performs preliminary analysis to create ground confidence maps before the actual mosaic stitching process. This preliminary action identifies valid ground areas and potential cut-line routes in advance, enabling the automated system to handle complex routing scenarios reliably during the subsequent stitching phase without requiring human intervention.
3Manufacturing precision
If cut lines are steered through ground areas to avoid three-dimensional objects, then visual accuracy is improved, but the complexity of determining preferred routes increases
Solution Approach 1:
The system performs preliminary analysis to create ground confidence maps that pre-identify valid ground areas and implicit cut-line routes. This preliminary action simplifies the subsequent stitching process by providing a ready-made guide for cut-line steering, reducing the computational complexity of route determination while maintaining high visual accuracy.
Solution Approach 2:
Ground confidence maps serve as an intermediary that encapsulates the complexity of ground versus non-ground discrimination. Rather than directly solving the complex problem of routing through three-dimensional objects, the system uses these maps as a simplified guide that automatically encodes the optimal cut-line paths through ground areas.
Data Source
AI summary
Systems and methods are disclosed for creating a mosaic image of two or more geo-referenced source images, the geo-referenced source images having the same orientation, based on a ground confidence map created by analyzing pixels of one or more of the geo-referenced source images, the ground confidence map having values and data indicative of particular geographic locations represented by the values, at least one of the values indicative of a statistical probability that the particular geographic locations represented by the values represents the ground; and using routes for steering mosaic cut lines based at least in part on the values indicative of the statistical probability that the particular geographic locations represented by the values represents the ground of the ground confidence map, such that the routes have an increased statistical probability of cutting through pixels representative of the ground versus routes not based on the ground confidence map.