Digital Image Noise Reduction Using Reference Dark Current Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Digital images produced by image sensors, such as CMOS sensors, are contaminated with noise like dark current, vertical patterns, and offset, which hinder quantitative analysis, especially in low signal-to-noise scenarios, and existing methods struggle to accurately estimate and remove these noises without temperature or exposure information.
Innovation Solution
A statistical analysis process using a reference digital image, which is free of vertical patterns and has been corrected for dark current, is applied to estimate and subtract dark current noise from target images, allowing for noise reduction without requiring temperature or exposure data, and further correcting for offset and vertical patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If dark current noise is present in digital images, then image quality deteriorates and quantitative analysis becomes inaccurate, but traditional noise reduction methods require temperature and exposure information that are not always available
Solution Approach 1:
The patent creates a reference digital image by processing a dark current digital image through de-correlation with a black digital image. This reference image copies the noise characteristics without requiring temperature or exposure metadata, enabling noise estimation through statistical analysis of the reference image alone.
Solution Approach 2:
The patent introduces a reference digital image as an intermediary that mediates between the noisy target image and the noise reduction process. The reference image, derived from dark current and black image data, serves as a template for estimating and subtracting noise without requiring direct measurement of temperature or exposure parameters.
2Ease of operation
If statistical analysis is applied to estimate noise magnitude, then noise reduction is achieved without temperature data, but the process requires additional processing steps and reference images
Solution Approach 1:
The patent performs preliminary processing by acquiring and processing dark current digital images and black digital images to create a reference digital image before analyzing the target image. This reference image is pre-computed and stored, allowing rapid noise estimation through statistical analysis without repeating the complex processing steps for each target image.
3Reliability
If dark current noise is removed using traditional methods, then noise is reduced, but the methods require temperature and exposure information that increases system complexity
Solution Approach 1:
The patent creates a reference digital image by processing a dark current digital image through de-correlation with a black digital image. This reference image copies the noise characteristics without requiring temperature or exposure metadata, enabling noise estimation through statistical analysis of the reference image alone.
Solution Approach 2:
The patent replaces the traditional mechanical approach of using temperature sensors and exposure time metadata with a statistical processing approach. Instead of relying on physical measurements of temperature and exposure, the system uses statistical analysis of reference images to estimate and remove noise, substituting physical measurement mechanisms with computational methods.
Data Source
AI summary
A target digital image is received from an image sensor. The image is contaminated by noise of unknown magnitude that is represented by a reference digital image. A process is applied that uses statistical analysis of the target digital image and of the reference digital image to estimate a magnitude of the noise for at least some pixels of the target digital image.