HDR Image Composition Using SDR Alignment Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The accuracy of detecting misregistration and moving objects in HDR image composition is reduced due to differences in tone characteristics caused by the gamma curve corresponding to HDR standards like ST2084, leading to a decrease in the quality of composite images.
Innovation Solution
An image processing apparatus and method that performs HDR development on multiple images captured under different conditions, generates composition information using images less influenced by tone processing from HDR development, and composites these images using SDR images or linear space images to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If HDR development with gamma curve (ST2084) is applied to captured images, then the brightness reproduction is improved to match human eye perception, but the tone characteristics become significantly different causing reduced accuracy in detecting misregistration and moving objects
Solution Approach 1:
The patent segments the image processing workflow into two distinct stages: (1) composition information generation using SDR images or linear space images that are not influenced by HDR tone processing, and (2) HDR composition using the generated composition information. This segmentation allows accurate detection of misregistration and moving objects in the first stage while maintaining HDR brightness reproduction in the second stage, thereby resolving the contradiction between brightness reproduction and detection accuracy.
Solution Approach 2:
The patent introduces SDR images or linear space images as intermediary elements that serve as a bridge between the captured images and the final HDR composite image. These intermediary images have stable tone characteristics suitable for accurate composition information generation, while the final HDR images maintain the desired brightness reproduction. The composition information generated from the intermediary images is then applied to composite the HDR images, thus resolving the contradiction.
2Illumination intensity
If composition information is generated using HDR images with tone processing, then the brightness reproduction is maintained, but the quality of composite image decreases due to reduced detection accuracy
Solution Approach 1:
The patent divides the processing into two segments: generating composition information from SDR/linear space images (ensuring high detection accuracy and composite image quality) and compositing HDR images using that composition information (maintaining brightness reproduction). This segmentation resolves the contradiction by assigning different optimization goals to different processing stages.
Solution Approach 2:
The patent performs preliminary action by generating composition information using SDR images or linear space images before performing the HDR composition. This preliminary step ensures accurate detection of misregistration and moving objects, which then guides the HDR composition process to produce high-quality composite images with proper brightness reproduction.
Data Source
AI summary
An image processing apparatus obtains a plurality of captured images captured by an image capturing unit under different conditions, performs HDR development of the plurality of captured images to obtain a plurality of HDR images, generates composition information for compositing the plurality of HDR images, using a plurality of images corresponding to the plurality of captured images, the plurality of images being not influenced by or less influenced by tone processing resulting from the HDR development, and composites the plurality of HDR images using the composition information.