Image Alignment via Location-Based Feature Matching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing computer vision methods struggle to accurately determine the 2D or 3D relationships between images of the same object taken at different times, especially due to seasonal variations, lighting changes, and changes in the urban landscape, leading to mismatched feature descriptors between corresponding feature points.

Innovation Solution

The method involves receiving feature points from two digital images, determining a translation offset by matching their locations without considering the feature descriptors, and applying this offset to graphical overlay objects, using algorithms like SIFT or SURF for feature detection, and employing a system to adjust the positions of these objects across images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If feature descriptors are used to match feature points between images, then feature matching accuracy is improved, but matching reliability deteriorates when images are taken at different times with seasonal variations, lighting changes, or urban landscape changes

Engineering Contradiction:

Improvefeature matching accuracyVSAvoidmatching reliability

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The matching process is segmented into two distinct stages: first matching feature point locations without descriptors to establish correspondence, then using descriptors to verify and refine the match. This segmentation allows the system to handle temporal variations robustly while maintaining feature-level matching accuracy when conditions permit.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Feature point locations serve as an intermediary between the unreliable feature descriptors and the final matching decision. By first establishing location-based correspondences and then using descriptors as a secondary verification mechanism, the system mediates between the conflicting requirements of accuracy and reliability under varying conditions.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Loss of information

If feature descriptors are compared for matching, then detailed feature information is utilized, but matching fails when seasonal variations, lighting changes, or urban landscape changes occur between image captures

Engineering Contradiction:

Improvefeature information utilizationVSAvoidmatching success rate

Core Design Contradiction:

Loss of informationVSReliability

Solution Approach 1:

The system performs preliminary matching based on feature point locations before applying the more sensitive descriptor comparison. This preliminary action establishes a set of candidate correspondences that are then verified using descriptors, allowing the system to retain detailed feature information while protecting against failures due to environmental changes.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The matching process applies descriptor comparison partially - only to feature points that have been preliminarily matched based on location. This partial application of descriptor analysis maintains robustness against environmental changes while still utilizing detailed feature information where it can be reliably applied.

Inventive Principle:

Principle #16Partial or excessive action

3Productivity

If traditional feature matching algorithms are used, then feature point correspondence is established, but translation offset calculation is inaccurate when images are taken at different times with environmental changes

Engineering Contradiction:

Improvefeature point correspondence establishmentVSAvoidtranslation offset accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system replaces the traditional single-step descriptor-based matching mechanism with a two-stage process that substitutes location-based preliminary matching for the initial correspondence establishment. This substitution makes the translation offset calculation more robust to environmental changes while maintaining productivity in establishing feature point correspondences.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS8971640B1Image alignment

Publication Date: 2015.03.03 GOOGLE LLC

AI summary

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for graphical overlay object translation. One of the methods includes receiving a first plurality of first feature points, each first feature point being associated with a plurality of feature descriptors and a respective location on the first digital image. The method includes receiving a second plurality of second feature points, each second feature point being associated with a plurality of feature descriptors and a respective location on the second digital image. The method includes determining the translation offset for the second digital image from the first digital image by matching locations of the first feature points with locations of the second feature points without regard to the feature descriptors.