Image Stitching Using Depth Parallax to Remove Ghosting
Find Innovative SolutionsGenerate 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
Engineering 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
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.
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.
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
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.
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.
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
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.
Data Source
Figure 1~2
Figure 3
Figure 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.