Recovering Saturated Pixel Values via Weighted Color Ratios
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current camera systems face challenges in capturing images with large dynamic ranges due to limited sensor dynamic range, leading to loss of information in shadows and highlights, and existing methods for reconstructing saturated pixels often result in color casts and contour artifacts.
Innovation Solution
An image reconstruction method that divides unsaturated pixels into color clusters based on specific pixel parameters, calculates weighting coefficients, and uses these to estimate saturated pixels, allowing for accurate reconstruction of type I, II, and III saturated pixels by combining information from unsaturated pixels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If the sensor dynamic range is increased to capture scenes with large dynamic range, then the loss of information in shadows and highlights is reduced, but the sensor full-well capacity and noise floor constraints make this difficult to achieve
Solution Approach 1:
The patent uses unsaturated pixels as intermediary elements to estimate and reconstruct saturated pixel values. By leveraging the correlation between R, G, and B channels in unsaturated regions, the method indirectly recovers information from saturated regions without requiring the sensor to physically capture the full dynamic range.
Solution Approach 2:
The patent creates a copy or estimate of saturated pixel information by using corresponding unsaturated pixels from the same or adjacent color channels. This allows the reconstruction of highlight and shadow details that would otherwise be lost due to saturation, effectively copying the missing information from reliable sources.
2Object-generated harmful factors
If traditional white balancing is applied to saturated pixels, then color cast is eliminated, but information loss in the saturated channels increases
Solution Approach 1:
The patent converts the harmful effect of saturation (which causes color cast) into a beneficial opportunity for information recovery. Instead of simply clipping saturated pixels, the method uses the saturation condition itself to identify regions where reconstruction is needed, and leverages the strong correlation between color channels to recover information that would otherwise be lost.
Solution Approach 2:
The patent changes the approach from direct white balancing (which eliminates color cast but loses information) to a reconstruction approach that estimates saturated pixel values based on unsaturated counterparts. This parameter change in the processing strategy allows simultaneous elimination of color cast and preservation of information through intelligent estimation.
3Device complexity
If simple clipping methods are used for saturated pixels, then processing complexity is reduced, but image quality and plausibility deteriorate due to contour artifacts and color casts
Solution Approach 1:
The patent segments the image processing task by first identifying saturated versus unsaturated pixels, then applying different processing strategies to each segment. Unsaturated pixels are used as reference data, while saturated pixels undergo reconstruction based on their unsaturated counterparts. This segmentation allows the method to focus computational resources only where needed, improving accuracy without excessive complexity.
Data Source
AI summary
Methods for recovering saturated pixel values in raw pixel data are described. Given an image with raw pixel values captured using sensors with a color filter array (CFA), image regions are classified according to how many color channels of the CFA are saturated. Values of saturated pixels are estimated based on weighted color ratios of unsaturated pixels and recovered pixels. Example methods for a Bayer CFA are provided.