Panoramic Image Stitching via Edge Extension Controller
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current display systems for panoramic images often suffer from stitching gaps due to projector misalignment, which affects the viewing experience by creating unaligned images at stitching boundaries.
Innovation Solution
A display system comprising N projectors and a controller that generates overlapping sub-images by copying and extending image edges, allowing adjacent sub-images to partially overlap, thus avoiding stitching gaps and improving alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If images are stitched side by side using conventional methods, then the panoramic image can be formed, but stitching gaps appear due to projector misalignment
Solution Approach 1:
The patent applies preliminary action by pre-extending the image data at the edges before stitching occurs. The controller extends the image data in the horizontal direction at both left and right edges, creating an extended image that includes overlapping regions. This preliminary extension ensures that when sub-images are stitched, the overlapping areas provide sufficient image data to cover any misalignment gaps, thus preventing stitching gaps before they can occur during the stitching process.
Solution Approach 2:
The patent changes the parameter of image data by extending it beyond the original boundaries. The controller generates extended image data that includes additional pixels beyond the original image edges, creating an extended image with increased width. This parameter change (extending image boundaries) provides extra image data in overlapping regions, ensuring that misalignment does not create visible gaps during stitching.
2Area of stationary object
If multiple projectors are used to form panoramic image, then the coverage area is increased, but alignment precision deteriorates due to positional deviation
Solution Approach 1:
The controller performs preliminary extension of image data at the edges before distribution to projectors. By extending the image data horizontally at both left and right edges and creating overlapping regions, the system ensures that each projector receives image data that extends beyond its designated display area. This preliminary action compensates for positional deviations, ensuring that even when projectors are misaligned, the overlapping extended regions provide sufficient image data to maintain seamless panoramic display across the increased coverage area.
3Reliability
If image edges are copied and extended, then stitching gaps are prevented, but image data processing complexity increases
Solution Approach 1:
The patent applies copying by duplicating the image data at the edges. The controller copies the image data in the horizontal direction at both left and right edges of the original image, creating extended image data that includes these copied regions. This copying process generates overlapping areas in the extended image that can be used to fill stitching gaps, ensuring high stitching quality while maintaining a relatively simple processing approach.
Data Source
AI summary
A display system for displaying a panoramic image and an operation method thereof are provided. The display system for displaying a panoramic image includes N projectors and a controller. The N projectors project N sub-images according to N image data and form a panoramic image. Each sub-image has a sub-image overlapping area, so that two adjacent sub-images of the N sub-images partially overlap, and that a first sub-image and an Nth sub-image of the N sub-images partially overlap. The controller copies a first edge image corresponding to a first edge of an original image and connects the first edge image to a second edge of the original image that is opposite to the first edge, so as to generate an extended image corresponding to the sub-image overlapping area. The controller slices an adjusted image containing the original image and the extended image to generate the N image data.


