Image Processing Device Raw Data Synthesis Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies fail to effectively combine the advantages of processing raw image data and full-color image data, leading to potential reduction in image quality due to inconsistent information handling before and after demosaicing.
Innovation Solution
An image processing device with a first processing unit for synthesizing raw image data by aligning corresponding pixels from multiple raw image data sets and a map generation unit for generating map information, followed by distinct image processing on demosaiced data to enhance noise reduction and visibility adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image processing is performed on raw image data before demosaicing, then noise reduction accuracy is improved by maintaining natural law information, but the ability to perform visibility-enhancing processing is reduced
Solution Approach 1:
The patent segments the image processing workflow into two distinct processing paths: one for raw image data (for noise reduction) and one for demosaiced full-color image data (for visibility enhancement). This allows each path to be optimized for its specific purpose without compromising the other.
Solution Approach 2:
The patent introduces an intermediary processing stage where noise reduction is performed on raw image data before demosaicing, and then the results are combined with processing on the demosaiced data. This intermediary step preserves natural law information while enabling subsequent visibility enhancements.
2Ease of operation
If image processing is performed on full-color image data after demosaicing, then visibility enhancement is improved, but noise reduction accuracy deteriorates due to loss of natural law information
Solution Approach 1:
The patent performs noise reduction processing as a preliminary action on raw image data before demosaicing occurs. This ensures that natural law information is preserved during the noise reduction process, achieving higher accuracy while still allowing subsequent visibility enhancements on the demosaiced data.
3Device complexity
If the same processing is executed on both raw image data and full-color image data, then device complexity is reduced through common circuit usage, but image quality deteriorates due to inability to combine advantages of different processing stages
Solution Approach 1:
The patent implements a processing system that can handle both raw image data and demosaiced full-color image data through a unified architecture. The same processing device performs different types of processing on different data types, combining the advantages of both processing paths while maintaining reasonable device complexity.
Data Source
AI summary
An image processing device includes: a first processing unit that detects corresponding pixels between reference image data selected from a plurality of pieces of raw image data and each of a plurality of pieces of comparison image data included in the raw image data, and synthesizes the reference image data and the comparison image data on the basis of the corresponding pixels to generate synthesized raw image data; a map generation unit that generates map information associating a pixel position of the synthesized raw image data with information derived from at least one piece of the raw image data; and a second processing unit that executes image processing different from that of the first processing unit on the demosaiced synthesized raw image data on the basis of the map information generated by the map generation unit.