Image Splicing via Primary and Partial Position Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image splicing technologies face issues when using a primary position parameter to project images, as it may not be applicable to the entire image, leading to undesirable registration and splicing effects due to uneven distribution of feature points across depth planes.
Innovation Solution
The method involves determining a primary position parameter for a current image based on an overlapped region, selecting a partially overlapped region with consistent depth planes, calculating a partial position parameter for each partial image region, and projecting the image again using both parameters to achieve improved registration and splicing effects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a primary position parameter is used to project the entire current image, then the projection process is simple and fast, but the registration accuracy and splicing effect deteriorate in regions where the primary parameter is inapplicable
Solution Approach 1:
The patent divides the image projection process into two stages: first projecting the entire image using a primary position parameter, then identifying and re-projecting specific partial image regions using partial position parameters. This segmentation allows the system to maintain high overall processing speed while achieving high registration accuracy in critical overlapping regions where the primary parameter may be inapplicable.
Solution Approach 2:
The patent applies different projection parameters to different regions of the image based on local requirements. The primary position parameter is used for regions where it provides adequate accuracy, while partial position parameters are specifically applied to overlapping regions where higher precision is needed. This local differentiation resolves the contradiction by optimizing both speed and accuracy in their respective appropriate regions.
2Device complexity
If a primary position parameter is used for the entire image, then the processing complexity is low, but the splicing quality deteriorates in overlapping regions
Solution Approach 1:
The patent segments the image into different processing zones: full-image regions processed with the primary position parameter and partial overlapping regions processed with partial position parameters. This segmentation strategy maintains low overall processing complexity while ensuring high splicing quality in the critical overlapping regions where feature point distribution may be uneven.
Solution Approach 2:
Instead of applying complex partial parameter processing to the entire image, the patent applies it only partially to specific overlapping regions where it is most needed. This partial action approach maintains acceptable processing complexity while significantly improving splicing quality in the regions where the primary parameter fails to provide adequate accuracy.
3Reliability
If feature points are unevenly distributed across depth planes, then the primary position parameter becomes inapplicable to certain regions, but using multiple partial parameters increases processing time
Solution Approach 1:
The patent segments the processing task by first performing a complete projection with the primary parameter, then identifying specific partial regions where feature point distribution indicates parameter inapplicability. This segmentation allows the system to maintain high reliability through targeted partial re-projection while minimizing additional processing time by avoiding unnecessary re-processing of regions where the primary parameter remains valid.
Solution Approach 2:
The system performs self-assessment by evaluating feature point distribution in overlapping regions to automatically determine where partial position parameters are needed. This self-service mechanism ensures parameter applicability reliability without requiring external intervention or excessive processing time, as the system intelligently identifies and processes only the regions that require enhanced accuracy.
Data Source
AI summary
Embodiments of this application disclose an image splicing method and apparatus, a terminal, and a storage medium thereof. The image splicing method includes determining a primary position parameter of a current image according to a first overlapped region in the current image; and projecting the current image to a projection canvas according to the primary position parameter to obtain an initial spliced image, the initial spliced image being obtained by splicing the current image and the neighboring image. The method further includes selecting a partially overlapped region from a second overlapped region of the initial spliced image; calculating a partial position parameter of a partial image region corresponding to the partially overlapped region in the current image; and projecting the current image to the projection canvas again according to the primary position parameter and the partial position parameter, to obtain a target spliced image.