RGB-IR Pixel Interpolation with Adaptive Weights for Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing RGB-IR image processing technologies suffer from infrared noise degradation and reduced visible resolution due to the spatial distribution of infrared-sensitive pixels, leading to structural and color artifacts, particularly in areas of high contrast, and are not suitable for conventional image processing formats like Bayer.
Innovation Solution
An interpolation technique that adjusts weights based on spatial uniformity and texture variations in the image, prioritizing the influence of neighboring pixels in uniform areas to refine image quality, including depollution processing, reconstruction of missing components, and formatting into a Bayer matrix.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If bilinear interpolation techniques are used for infrared depollution and reconstruction, then processing complexity is reduced, but image quality deteriorates due to structural and color artifacts in high contrast areas
Solution Approach 1:
The patent applies local quality by using adaptive interpolation weights that vary spatially across the image. Instead of uniform bilinear interpolation, the method calculates weights based on local image characteristics (gradient magnitude, texture complexity) to preserve edges and reduce artifacts in high contrast areas while maintaining smooth interpolation in uniform regions.
Solution Approach 2:
The patent introduces dynamics by making interpolation weights adaptive rather than static. The weights are dynamically calculated based on local image properties such as gradient magnitude and texture measures, allowing the interpolation process to respond to local variations in image content and avoid artifacts in edge regions.
2Area of stationary object
If infrared-sensitive pixels are distributed among red, green, and blue pixels in an RGB-IR matrix, then size is reduced, but visible resolution deteriorates due to the presence of dedicated infrared pixels
Solution Approach 1:
The patent uses copying by reconstructing the visible components at infrared pixel locations through interpolation from neighboring visible pixels. This creates virtual copies of the visible information that would otherwise be blocked by the infrared pixel structure, effectively increasing the usable visible pixel density without changing the physical sensor layout.
Solution Approach 2:
The patent applies parameter changes by transforming the raw RGB-IR pixel data into reconstructed visible components through mathematical interpolation. This changes the parameter representation from direct sensor readings to processed estimates, allowing the system to overcome the physical limitations of the interleaved pixel layout.
3Reliability
If infrared wavelength band is not filtered for visible light-sensitive pixels, then sensitivity is improved, but infrared noise degradation increases
Solution Approach 1:
The patent extracts and removes infrared noise from visible pixel signals through interpolation-based depollution. By calculating what the visible signal should be based on neighboring pixels and local image characteristics, the method separates and removes the infrared contamination component, preserving pixel sensitivity while eliminating noise.
Solution Approach 2:
The patent uses interpolated visible components from neighboring pixels as an intermediary to estimate and remove infrared noise. This intermediary representation serves as a reference for what the clean visible signal should look like, allowing the system to subtract the infrared contamination without losing sensitivity.
4Productivity
If classic bilinear interpolation is used for Bayer formatting, then processing speed is maintained, but image quality deteriorates due to artifacts in transition zones
Solution Approach 1:
The patent applies local quality by using adaptive interpolation weights that vary spatially across the image. Instead of uniform bilinear interpolation, the method calculates weights based on local image characteristics (gradient magnitude, texture complexity) to preserve edges and reduce artifacts in high contrast areas while maintaining smooth interpolation in uniform regions.
Solution Approach 2:
The patent introduces dynamics by making interpolation weights adaptive rather than static. The weights are dynamically calculated based on local image properties such as gradient magnitude and texture measures, allowing the interpolation process to respond to local variations in image content and avoid artifacts in edge regions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The process for processing a matrix of pixels (RGBIR_RAW) each containing an original red (R), green (G), blue (B), or infrared (IR) component, includes at least one interpolation of an interpolated component (R) different from the original component (IR) of a pixel of interest (P) from the components of a group (KER) of pixels neighboring the pixel of interest (P). The interpolation includes: - a calculation of the sum of the components of reference pixels (P1, P2) weighted by a weight respectively assigned, the reference pixels (P1, P2) being pixels of the group (KER) having the same original component (R) as the interpolated component (R), - an evaluation of a spatial uniformity of an environment (P11, ..., P14 ; P21, ..., P24), within the group (KER), of each reference pixel (P1, P2), - a calculation of the weights assigned to the reference pixels (P1, P2) to normalized values and proportional to the respective spatial uniformity.