CMOS Sensor RTS Noise Detection via Pixel Value Grouping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Image sensors, particularly CMOS sensors, face challenges with reduced sensitivity and increased noise due to miniaturization, including blinking defect noise like random telegraph signal (RTS) noise, which affects pixel value consistency and requires effective detection and correction methods.
Innovation Solution
An image processing apparatus and method that classify pixel values for each read-out circuit to create pixel value groups, determine blinking defect noise, specifically RTS noise, by analyzing distribution patterns, and correct noise levels using candidate values and random noise models to isolate and correct RTS noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If pixel size is miniaturized to reduce sensor area and increase pixel density, then the aperture ratio and sensitivity are improved, but noise increases including blinking defect noise
Solution Approach 1:
The patent segments the analysis by dividing pixels into groups based on their read-out circuits, and further segments the noise analysis by comparing pixel values across multiple imaging shots. This segmentation allows identification of blinking defect noise specifically affecting certain pixel groups while maintaining the miniaturized pixel structure benefits.
Solution Approach 2:
The patent employs periodic action by capturing multiple imaging shots and analyzing pixel value variations across these periodic measurements. This allows detection of blinking defect noise that manifests as temporal variations, distinguishing it from static defects and random noise.
2Area of stationary object
If sharing pixel structure is used to reduce read-out circuit area, then aperture ratio is improved, but blinking defect noise attributed to read-out circuits increases
Solution Approach 1:
The patent segments pixels into groups based on shared read-out circuits, allowing identification of which specific read-out circuits exhibit blinking defect noise. This enables targeted correction while maintaining the area-efficient sharing structure.
Solution Approach 2:
The patent applies partial correction by identifying and correcting only those pixel values affected by blinking defect noise, rather than applying blanket correction to all pixels. This excessive action ensures that noise-corrected values are used even when some correction may be applied to normal variations.
3Measurement precision
If conventional RTS noise detection is used by subtracting average pixel values, then detection capability is achieved, but false detection and over-correction occur
Solution Approach 1:
The patent segments the detection process by analyzing pixel values in groups rather than individually, and by comparing distributions across multiple shots. This segmentation reduces false detection by requiring consistent patterns across multiple measurements rather than relying on single-shot anomalies.
Solution Approach 2:
The patent employs feedback by iteratively comparing pixel value distributions against expected patterns and adjusting detection thresholds based on observed variations. This feedback mechanism refines the detection process to distinguish true blinking defect noise from normal pixel variations.
4Measurement precision
If multiple imaging shots are used to detect blinking defect noise, then detection accuracy is improved, but processing time increases
Solution Approach 1:
The patent segments the large dataset from multiple imaging shots into smaller pixel value groups organized by read-out circuit. This segmentation enables efficient processing by analyzing smaller subsets simultaneously rather than processing all pixel values from all shots uniformly.
Solution Approach 2:
The patent applies partial processing by focusing analysis only on pixel groups that exhibit suspicious variations, rather than uniformly processing all pixel groups from multiple shots. This reduces processing time while maintaining detection accuracy for affected regions.
Data Source
AI summary
An image processing apparatus includes: a pixel value group creation unit configured to create, for a plurality of pieces of image data generated by an image sensor including: a plurality of pixels arranged two-dimensionally to receive light from outside and generate a signal corresponding to an amount of the received light; and a plurality of read-out circuits shared by a predetermined number of pixels and configured to read out the signal as pixel values, a plurality of pixel value groups by classifying the pixel values for each of the plurality of read-out circuits; and a noise determination unit configured to determine whether blinking defect noise occurs in each of the plurality of pixel value groups, based on distribution of the pixel values of each of the plurality of pixel value groups created by the pixel value group creation unit.