Noise Reduction Device Using Blackout Image Thresholding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional noise reduction methods for image data from electronic cameras during long exposure increase random noise and require equal charging times for normal and blackout images, leading to prolonged blackout image capture times that hinder photography opportunities and increase memory consumption.
Innovation Solution
A noise reduction device and method that captures image data and blackout image data with a shading mechanism, extracts specific noise components, and processes them to suppress high spatial frequency components and dark current offsets, allowing for shorter blackout image charging times without significantly increasing random noise, thereby reducing fixed pattern noise effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the conventional device subtracts blackout image data from normally taken image data to remove fixed pattern noise, then fixed pattern noise is reduced, but random noise is increased due to reversed mode addition
Solution Approach 1:
The patent extracts only the specific noise component from blackout image data by thresholding (extracting values equal to or larger than a predetermined value) rather than using the entire blackout image. This extraction isolates the fixed pattern noise while excluding most random noise, allowing effective noise reduction without introducing reversed random noise artifacts.
Solution Approach 2:
The patent applies different processing to different parts of the blackout image data: low-level pixels (below threshold) are suppressed to remove random noise, while high-level pixels (above threshold) are retained to preserve fixed pattern noise information. This local differentiation enables selective noise component extraction.
2Manufacturing precision
If the charging time of blackout image data is made equal to that of normally taken image data to match fixed pattern noises, then fixed pattern noise removal accuracy is improved, but shooting opportunities are missed due to prolonged blackout time
Solution Approach 1:
The patent changes the parameter of blackout image charging time to be shorter than that of normally taken image data. By combining this shortened time with signal amplification and selective noise component extraction through thresholding, the system achieves effective fixed pattern noise removal without requiring equal charging times, thus reducing time loss.
Solution Approach 2:
The patent creates a simplified copy of the blackout image data by extracting only the specific noise component above the threshold. This copied noise component contains the essential fixed pattern information needed for subtraction, allowing the use of shortened charging times while maintaining removal effectiveness.
3Manufacturing precision
If the blackout image data is temporarily stored in memory to process fixed pattern noise, then noise reduction processing is enabled, but memory consumption is increased
Solution Approach 1:
The patent extracts only the necessary noise component from blackout image data through thresholding, creating a compact representation that contains only the fixed pattern noise information. This extracted component requires minimal memory storage compared to storing the entire blackout image, thus reducing memory consumption while maintaining noise reduction capability.
Data Source
AI summary
The noise reduction device includes an image storage unit, a blackout image processing unit, and a noise processing unit. The image storage unit captures image data obtained by imaging a field with an image sensor, and stores the image data therein. The blackout image processing unit captures blackout image data obtained by imaging by the image sensor that is shaded, and extracts a specific noise component of the blackout image data. The noise processing unit reduces a noise in the image data based on the specific noise component of the blackout image data.