Image Deblurring via Dual-Exposure Segmentation and Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies fail to effectively deblur images captured with improperly set camera parameters or in challenging lighting environments, resulting in blurred images that are not sharp enough for optimal attractiveness and utility.
Innovation Solution
A method that records both short-exposed and normal-exposed pixels, using the former to enhance the sharpness of the latter by calculating a correlation coefficient and applying deblurring models or filters to produce a clearer image, while also updating camera parameters based on image sharpness values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If normal exposure is used for full resolution imaging, then image coverage is complete, but image sharpness deteriorates in challenging lighting environments
Solution Approach 1:
The image sensor is divided into two regions: a first region that captures short-exposed pixels at a higher frame rate and a second region that captures normal-exposed pixels at a normal frame rate. This segmentation allows different parts of the image to be optimized for different purposes, with the short-exposed region providing sharpness information and the normal-exposed region providing full dynamic range coverage.
Solution Approach 2:
A deblurring model is introduced as an intermediary computational process that takes both short-exposed and normal-exposed pixels as input and produces a deblurred normal-exposed image. This intermediary model acts as a bridge that combines the advantages of both exposure types to achieve sharp images with full dynamic range.
2Manufacturing precision
If short-exposed pixels are recorded at higher frame rate, then image sharpness is improved, but device complexity increases
Solution Approach 1:
The system performs preliminary capture of short-exposed pixels at high frame rate to establish a sharp reference image before the final image processing. This preliminary action provides the necessary sharpness information that is then used by the deblurring model to enhance the normal-exposed image, separating the sharpness capture function from the final image synthesis.
Solution Approach 2:
The deblurring model uses feedback from both short-exposed and normal-exposed pixels to iteratively optimize the deblurred image output. The model learns from the correlation between the sharp short-exposed regions and the full-range normal-exposed regions, adjusting its parameters to produce the best possible sharp image while managing processing complexity through efficient algorithm design.
Data Source
AI summary
For deblurring an image, a method records both short-exposed pixels at a higher frame rate for a short-exposure region and normal-exposed pixels at a normal frame rate for full resolution. In addition, the method deblurs a normal-exposed image as a function of the short-exposed pixels and the normal-exposed pixels.