Image Stabilization Using Predicted Feature Point Positions

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image stabilization schemes are susceptible to external impacts and image quality degradation, and struggle to correct motion accurately when moving objects are present, especially when motion deviates from the tracking region or exceeds the tracking window size.

Innovation Solution

An image stabilizing apparatus and method using a sensor fusion scheme that includes a feature point extractor, movement amount detector, movement position predictor, corresponding position detector, comparator, and image stabilizer to predict and correct image distortion caused by camera movement, utilizing both image and motion sensors to determine accurate feature point positions and correct image distortion within a predetermined range.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If image processing methods are used to correct motion, then image stabilization can be achieved, but when motion deviates from tracking region or exceeds tracking window size, the motion correction becomes difficult

Engineering Contradiction:

Improvemotion detection accuracyVSAvoidhandling large motion deviations

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent introduces an intermediary comparison mechanism that validates feature point matching results against predicted positions. When motion exceeds tracking window size, the system uses a comparator to detect mismatches and switches to alternative correction methods, thereby extending the system's adaptability to large motion deviations while maintaining measurement precision for normal conditions

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent implements dynamic adjustment of tracking window size and feature point extraction parameters based on detected motion magnitude. When motion deviates from the tracking region, the system dynamically expands the search area and adjusts matching parameters, enabling the image stabilization system to adapt to varying motion conditions while maintaining accurate motion correction

Inventive Principle:

Principle #15Dynamics

2Manufacturing precision

If feature point matching is used for image stabilization, then image correction can be performed, but when moving objects exist in captured images, accurate correction cannot be achieved

Engineering Contradiction:

Improveimage correction accuracyVSAvoidinterference from moving objects

Core Design Contradiction:

Manufacturing precisionVSObject-affected harmful factors

Solution Approach 1:

The patent implements a feedback mechanism where the comparator continuously compares predicted feature point positions with actually detected positions. When moving objects cause mismatches, the feedback loop detects the discrepancy and excludes invalid feature points from correction calculations, thereby maintaining high image correction accuracy even in the presence of moving objects

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent dynamically changes parameters such as feature point selection criteria, matching thresholds, and tracking window parameters based on scene analysis. When moving objects are detected through parameter mismatches, the system adjusts parameters to focus on stationary features, thereby eliminating interference from moving objects and achieving accurate image correction

Inventive Principle:

Principle #35Parameter changes

3Reliability

If external impact or image quality degradation occurs, then image stabilization becomes unreliable, but traditional methods cannot handle these conditions

Engineering Contradiction:

Improvestabilization performanceVSAvoidexternal impact and quality degradation

Core Design Contradiction:

ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent performs preliminary prediction of feature point positions based on motion detection before actual feature point matching. This preliminary action creates expected position references that can be used to validate matching results, thereby maintaining reliable stabilization performance even when external impact or image quality degradation occurs during capture

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent prepares multiple feature point candidates and alternative matching strategies in advance to cushion against the effects of external impact and quality degradation. When degradation occurs, the pre-prepared alternatives can be activated to maintain stabilization reliability, preventing the harmful effects from compromising image quality

Inventive Principle:

Principle #11Beforehand cushioning (Prior cushioning)

Data Source

PatentUS9854171B2Image stabilizing apparatus and method based on a predicted movement position

Publication Date: 2017.12.26 HANWHA VISION CO LTD

AI summary

Provided are an image stabilizing apparatus and method thereof. The image stabilizing apparatus performs image stabilization by using both an image sensor and a motion sensor. Image distortion and movement are stably corrected by using both the position of a feature point, which is extracted by the image sensor and image processing, and the movement position of the feature point, which is predicted by the motion sensor.