Image Stitching Using Depth Parallax to Remove Ghosting

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image stitching methods face challenges in handling images captured from different viewpoints, leading to ghosting due to parallax, especially when backgrounds are not uniform, and require extensive calculations for intermediate viewpoint images.

Innovation Solution

The method involves finding the overlap region between two images, obtaining a depth image to calculate precise parallax, and stitching the images using this depth information, reducing the need for recalculating intermediate viewpoints and simplifying the process by using a depth camera or calculating parallax directly.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If images are obtained from different viewpoints and stitched together, then wide-view images can be captured, but ghosting occurs due to parallax between different viewpoints

Engineering Contradiction:
Improveability to capture images from different viewpointsVSAvoidghosting artifact in stitched image
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent performs preliminary action by calculating the parallax map before the actual stitching process. The parallax map pre-calculates the displacement vectors for each pixel based on depth information, allowing the stitching algorithm to compensate for parallax effects in advance and avoid ghosting artifacts during the stitching operation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary element - the parallax map - between the source images and the final stitched image. This parallax map serves as a mediator that contains depth information and guides the stitching process, enabling the system to handle images from different viewpoints without producing ghosting artifacts.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If intermediate viewpoint images are generated to deghost the stitched image, then parallax ghosts are reduced, but enormous calculations are required and stitching time increases significantly

Engineering Contradiction:
Improveghost reduction in stitched imageVSAvoidimage stitching time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent extracts only the essential information needed for ghost reduction - the parallax map containing depth information - without generating complete intermediate viewpoint images. This extraction approach obtains the necessary parallax compensation data while avoiding the computationally expensive process of generating and processing multiple intermediate images, thus reducing stitching time significantly.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies partial action by calculating and using only the parallax information for the overlap region rather than processing entire intermediate viewpoint images. This selective approach focuses computational resources only on the necessary parallax calculation for ghost reduction, avoiding the excessive calculations required by methods that generate complete intermediate images.

Inventive Principle:
Principle #16Partial or excessive action

3Manufacturing precision

If camera calibration is performed to obtain mapping relation, then image stitching can be achieved, but calibration is difficult in computer field and backgrounds must be set as single background

Engineering Contradiction:
Improvemapping relation accuracyVSAvoidcalibration difficulty
Core Design Contradiction:
Manufacturing precisionVSEase of manufacture

Solution Approach 1:

The patent substitutes the mechanical calibration process with a computational approach using depth maps and parallax calculations. Instead of physically calibrating camera parameters and assuming single backgrounds, the system uses depth information to directly compute parallax maps, eliminating the need for complex calibration procedures and background assumptions while achieving accurate mapping relations.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP2328125B1Image splicing method and device
Publication Date: 2013.01.16 HUAWEI DEVICE CO LTD
  • EP2328125B1 patent drawingFigure 1~2
  • EP2328125B1 patent drawingFigure 3
  • EP2328125B1 patent drawingFigure 4

AI summary

The present invention relates to an image processing technology, and discloses an image stitching method and apparatus to solve the problem of severe ghosting of an image stitched in the prior art. In the embodiments of the present invention, the overlap region of two images is found, a depth image of the overlap region is obtained, and the two images are stitched together according to the depth image. In the stitching process, the 3-dimensional information of the images is obtained by using the depth image to deghost the image. The method and apparatus under the present invention are applicable to multi-scene videoconferences and the occasions of making wide-view images or videos.