Eigenspace Projection for Dynamic Viewpoint Change Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional automatic image processing techniques are ineffective for dynamic image scenes from mobile platforms or cameras with changing viewpoints, such as UAVs, due to image transformations from differing viewpoints, leading to high human error and time consumption.

Innovation Solution

An image processing method using Principal Component Analysis (PCA) with Singular Value Decomposition (SVD) to create a model space for eigenspace projection, allowing for change detection by comparing distances between new and reference points, with viewpoint identification based on the mobile platform's position, and using a distance-based threshold to determine significant changes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional automatic image processing techniques are used for static viewpoints, then processing speed is improved, but they cannot handle dynamic scenes from mobile platforms due to image transformations

Engineering Contradiction:

Improveprocessing speedVSAvoidcapability to handle dynamic scenes

Core Design Contradiction:

ProductivityVSAdaptability or versatility

Solution Approach 1:

The system adapts to dynamic viewing conditions by continuously adjusting the image processing approach based on detected viewpoint changes. The method dynamically determines whether images are from static or mobile viewpoints and applies appropriate processing techniques, making the system versatile for both static and dynamic scene monitoring

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The invention changes the processing parameters based on the viewpoint characteristics. For mobile platform images, it transforms the approach from direct pixel comparison to eigenspace projection after detecting viewpoint changes, allowing the system to handle transformed images while maintaining processing efficiency

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If human operators visually monitor video feeds for change detection, then detection accuracy is improved, but time consumption and cost increase significantly

Engineering Contradiction:

Improvedetection accuracyVSAvoidtime consumption

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The invention replaces the mechanical human visual monitoring system with an automated computer-based image processing system. It uses eigenspace projection and automatic change detection algorithms to achieve accurate detection without human intervention, eliminating time consumption and cost associated with manual monitoring while maintaining or improving detection accuracy

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system introduces an intermediary processing layer (eigenspace projection) between raw image data and change detection. This intermediary representation allows automatic comparison and detection of changes while filtering out irrelevant variations, enabling accurate automated detection that replaces human monitoring

Inventive Principle:

Principle #24Intermediary (Mediator)

3Device complexity

If images from differing viewpoints are compared directly, then processing simplicity is maintained, but image transformations cause false detections and reduce accuracy

Engineering Contradiction:

Improveprocessing simplicityVSAvoidchange detection accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The system performs preliminary viewpoint analysis before change detection by determining whether images are from static or mobile viewpoints. This preliminary action allows the system to prepare appropriate processing methods in advance, preventing false detections caused by uncorrected viewpoint transformations while maintaining processing efficiency through pre-planned approaches

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentEP2502201B1Image processing for change detection

Publication Date: 2020.01.08 BAE SYSTEMS PLC

Figure 1
Figure 2~3
Figure 4~5

AI summary

An image processing method and apparatus, comprising: capturing image data from a sequence of images of a scene from different but repeatable viewpoints (14a-f), e.g. different positions on a repeatable surveillance path; associating captured images with a respective viewpoint identification;tracking a region of interest (10) in the sequence of images;performing an eigenspace projection for the respective images to calculate respective points (22a-f) in the eigenspace projection;comparing the calculated points (22a-f) to points previously obtained (20a-f) for a reference image for substantially the same viewpoints (14a-f);wherein a change between a calculated point (22f) and a corresponding previously obtained point (20f) indicates a change in the image content of the region of interest (10) between the corresponding captured image and the corresponding reference image for the substantially same viewpoint (14f).