Fixed Pattern Noise Removal Using Column Compensation Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for removing fixed pattern noise in image sensors require storing entire optical black pictures, leading to high memory consumption and computational complexity due to non-uniform pixel sampling and hardware architecture deviations.
Innovation Solution
Calculating compensation values for each pixel column, allowing for the compensation of active pixel values without storing the entire optical black picture, thereby reducing memory usage and computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the conventional method of storing entire optical black pictures is used for FPN removal, then the fixed pattern noise can be effectively removed, but the memory consumption increases significantly
Solution Approach 1:
The patent extracts only the essential information needed for FPN removal by calculating compensation values for each pixel column based on optical black pictures, rather than storing and processing complete optical black pictures. This extraction approach retains the noise removal effectiveness while dramatically reducing memory requirements.
Solution Approach 2:
Instead of storing the complete optical black picture, the patent creates a simplified representation (copy) in the form of compensation values for each pixel column. These compensation values capture the essential FPN characteristics without requiring storage of the full image data.
2Reliability
If the conventional method of storing and processing optical black pictures is used, then FPN can be removed, but the computational complexity increases
Solution Approach 1:
The patent segments the FPN removal process into two independent stages: (1) calculating compensation values for each pixel column from optical black pictures, and (2) applying these pre-calculated values to real images. This segmentation reduces computational complexity by avoiding the need to process complete optical black pictures during normal operation.
Solution Approach 2:
The patent performs preliminary calculation of compensation values for each pixel column using optical black pictures before actual image processing. This preliminary action allows the complex FPN removal computation to be done once during calibration, rather than repeatedly during normal image capture operations.
3Ease of manufacture
If non-uniform pixel sampling is used in the image sensor, then the hardware architecture is simplified, but fixed pattern noise is introduced
Solution Approach 1:
The patent changes the parameter representation from complete pixel values to compensation values for each pixel column. By working with these transformed parameters, the system can correct FPN introduced by non-uniform sampling without requiring uniform sampler design, thus maintaining hardware simplicity while eliminating the harmful noise effect.
Data Source
AI summary
A fixed pattern noise removal method for an image sensor includes steps of calculating each compensation value corresponding to each pixel column according to a plurality of compensation pixel values of the each pixel column; and sampling a plurality of active pixel values in an active pixel area and compensating the plurality of active pixel values according to the each compensation value of the each pixel column, to generate a plurality of compensated active pixel values. A plurality of active pixels sense light to generate the plurality of active pixel values.