Dynamic Defect Pixel Correction Using Spatially Arranged Exposures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing techniques are inadequate in dynamically detecting and correcting defect pixels in image data from CMOS image sensors, particularly in wide dynamic range modes, as they often fail to preserve image quality and local edges.
Innovation Solution
The system employs a defect pixel correction method that utilizes spatially arranged exposures with different subsets of pixels having distinct exposure times, dynamically adjusts neighbor pixel values based on exposure ratios, computes neighborhood statistics, and replaces defective pixel values with central tendency measures or averages from neighbor pixels, depending on whether they are in flat or non-flat regions, while inhibiting replacement under certain conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional dynamic pixel correction techniques are used to detect and replace defective pixels, then defect pixels can be corrected, but image quality and local edges are degraded
Solution Approach 1:
The patent applies local quality by analyzing neighborhood statistics (mean and standard deviation) of pixels surrounding the defective pixel to determine whether to replace it. The decision to replace a defective pixel depends on whether its value falls outside the range defined by mean ± 3σ. This localized statistical analysis preserves local image characteristics and edges while correcting defects, avoiding blanket replacement that would degrade image quality.
2Device complexity
If static pixel correction approaches are used, then processing is simple, but intermittent defects cannot be mitigated
Solution Approach 1:
The patent implements dynamic pixel correction by using neighborhood statistics computed from actual image data to determine whether each pixel is defective. Unlike static correction that uses fixed thresholds, this approach dynamically adapts to local image characteristics by calculating mean and standard deviation for each pixel's neighborhood, enabling detection of intermittent defects while maintaining processing efficiency through standardized statistical operations.
3Reliability
If defective pixels are replaced with fair approximations, then defect pixels are corrected, but image resolution and quality are reduced
Solution Approach 1:
The patent changes the parameter used for replacement by using a statistical approach (mean ± 3σ) rather than simple approximation. By computing neighborhood statistics and replacing only pixels that fall outside this statistically determined range, the method preserves image resolution and quality while correcting defects. The replacement value is derived from the neighborhood mean, which adapts to local image characteristics rather than using fixed approximation values.
Data Source
AI summary
Various technologies described herein pertain to defect pixel correction for image data collected by a pixel array of an image sensor with spatially arranged exposures. The pixel array includes a first subset of pixels having a first exposure time and a second subset of pixels having a second exposure time. An exposure ratio (ratio of first exposure time to second exposure time) is received. A value of at least a particular neighbor pixel of a given pixel from the image data is adjusted based upon the exposure ratio. Neighborhood statistics for the given pixel from the image data are computed based on values of neighbor pixels of the given pixel from the image data as adjusted. Whether the value of the given pixel is defective is detected based on the neighborhood statistics. The value of the given pixel is replaced when detected to be defective to output modified image data.