Image Stitching Warp Field Computation for Parallax Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cameras with limited field of view struggle to capture panoramic scenes accurately due to parallax artifacts caused by camera motion and scene changes, making it difficult to perfectly align overlapping images using a single transform, especially in hand-held or head-mounted scenarios.
Innovation Solution
An image stitching process that computes a warp field with displacement vectors for a tessellation of points in the overlapping region, using multi-threading and dynamic programming to efficiently align and stitch images, accounting for camera motions and scene changes, thereby reducing parallax artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single transform is used to align overlapping images, then the alignment process is simple and fast, but parallax artifacts appear and alignment precision deteriorates due to camera motion and scene changes
Solution Approach 1:
The patent divides the overlapping region into multiple local patches and computes separate displacement vectors for each patch. This segmentation allows the system to handle local variations in camera motion and scene changes independently, improving alignment precision without requiring a single complex global transform.
Solution Approach 2:
The patent uses dynamic programming to compute displacement vectors that adapt to local conditions in the overlapping region. Instead of applying a static single transform, the system dynamically adjusts the alignment parameters for each local patch based on actual image content and motion characteristics, resolving the contradiction between simplicity and precision.
2Area of stationary object
If multiple photographs are captured with overlapping fields of view to create panoramic images, then the field of view coverage is improved, but computational resources and processing time increase
Solution Approach 1:
The patent segments the warping process into discrete patches within the overlapping region, computing displacement vectors only where needed. This reduces the overall computational burden compared to processing entire images, enabling efficient handling of multiple photographs while maintaining wide field of view coverage.
Solution Approach 2:
The patent applies warping and alignment operations selectively only to the overlapping regions where they are most needed, rather than processing entire images uniformly. This local approach reduces computational resources while preserving the ability to create panoramic images from multiple photographs with varying overlap.
3Measurement precision
If warping operations are applied to align images with different camera motions, then alignment accuracy improves, but computational complexity and processing time increase
Solution Approach 1:
The patent divides the warping operation into discrete patches, computing displacement vectors for each patch independently using dynamic programming. This segmentation enables parallel processing and reduces the overall processing time while maintaining high alignment accuracy through localized warping that accounts for different camera motions in different regions.
Data Source
AI summary
A computing device is described which has a memory holding at least two input images depicting different parts of a panoramic scene, the images having been captured by a user moving the camera by hand to capture the panorama. The computing device has an image stitching component configured to identify, at a processor, a region of overlap between the at least two images and to calculate a displacement vector for each of a plurality of warp points in the region of overlap. The image stitching component is arranged to warp a second one of the at least two images using the warp points; and to join the warped second image to the first image.