Panorama Image Stitching Using ROI-Based HDR Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating panorama images struggle with accurately matching and stitching multiple images to produce high-quality, high-dynamic-range panorama content, often resulting in noticeable brightness differences and reduced matching success rates.
Innovation Solution
The proposed solution involves a method and apparatus that apply High Dynamic Range (HDR) processing to selected Regions Of Interest (ROIs) from multiple images, followed by stitching based on matching coordinates, using techniques like tone mapping and contrast limited adaptive histogram equalization to enhance image contrast and correlation, thereby improving the matching success rate and minimizing brightness differences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional image matching and stitching methods are used to generate panorama images, then the stitching process can be completed, but the matching success rate is low and brightness differences are noticeable between stitched images
Solution Approach 1:
The patent applies HDR processing to overlap regions before performing image matching and stitching. By pre-processing the images to equalize brightness and dynamic range in the overlapping areas, the system improves matching accuracy and reduces visible brightness differences in the final panorama, thereby resolving the contradiction between matching success rate and brightness consistency.
2Manufacturing precision
If HDR processing is applied to entire images before stitching, then brightness consistency is improved, but processing time and computational complexity increase significantly
Solution Approach 1:
The patent divides the images into overlap regions and non-overlap regions, applying HDR processing only to the overlap regions where matching is performed. This segmentation approach maintains brightness consistency in the critical stitching areas while avoiding the computational overhead of processing entire images, thus resolving the contradiction between brightness consistency and processing time.
Solution Approach 2:
The patent applies different processing qualities to different regions: HDR processing is applied locally to overlap regions to ensure brightness consistency for matching, while non-overlap regions remain unprocessed. This local quality approach optimizes processing efficiency by concentrating computational resources only where they are most needed for the stitching operation.
3Area of stationary object
If multiple images are stitched together to create wide-angle panorama content, then the field of view is expanded, but matching accuracy decreases due to varying lighting conditions and brightness differences
Solution Approach 1:
The patent performs HDR processing on overlap regions before matching to pre equalize brightness and dynamic range variations caused by different lighting conditions. This preliminary brightness normalization enables accurate feature matching across images with vastly different exposure levels, allowing the system to stitch multiple images into wide-angle panoramas while maintaining high matching accuracy.
Data Source
AI summary
There are provided an apparatus and a method for generating a panorama image. The apparatus includes: a Region Of Interest (ROI) selecting block for receiving a plurality of images, and outputting ROI images by selecting an ROI of each of the plurality of images; a High Dynamic Range (HDR) processing block for performing HDR processing on the ROI images; and a panorama image generating block for generating a panorama image by stitching the plurality of images using a matching coordinate of the ROI images on which the HDR processing is performed. The HDR processing block performs the HDR processing on the panorama image.