X-Droop Correction in Solid State Image Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing solid state image sensors face challenges in correcting x-droop noise due to differences in column ground plane voltage and process variations, as previous methods fail to account for temporal noise associated with ADC circuitry and are limited to specific readout methodologies, leading to residual noise and introduction of new Vertical Fixed Pattern noise components.
Innovation Solution
A method that uses neighboring pixels to provide averaged noise reduction, averaging samples both in time and across columns to subtract x-droop noise from visible rows, reducing temporal and fixed noise contributions, and incorporating an optional global average factor to maintain image mean.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If column FPN correction is performed using only correction values from the same column, then the correction mechanism is simple, but temporal noise contributions from ADC circuitry remain and new Vertical Fixed Pattern noise components are introduced
Solution Approach 1:
The patent combines correction values from multiple columns by averaging neighboring pixel values horizontally. Instead of using only the same column for correction, the system averages correction data across adjacent columns to create a more comprehensive correction profile that accounts for temporal noise and reduces the introduction of new fixed pattern noise components.
Solution Approach 2:
The patent extends the correction mechanism from a single-column approach to a multi-column approach by incorporating horizontal neighbor information. This adds a spatial dimension to the correction process, allowing the system to leverage correction data from multiple columns simultaneously to improve noise reduction effectiveness.
2Measurement precision
If sample and hold readout methodology is used with CDSSIG pulse, then reset noise contribution is removed, but temporal noise from comparator and kTC noise are stored and added to pixels
Solution Approach 1:
The patent converts the harmful temporal noise components (comparator noise and kTC noise) into beneficial correction data by measuring them during the readout process. These noise components are captured in the correction values from masked pixels and then used to correct the actual image pixels, thereby eliminating their harmful effects.
Solution Approach 2:
The patent creates a copy of the noise characteristics by measuring them in optically masked pixels that are identical to the visible pixels but do not contribute to the image. This copy of the noise profile is then applied to correct the visible pixels, effectively removing temporal noise contributions.
3Measurement precision
If continuous time system with auto-zero operation is used, then black value is corrected, but noise from source follower (RTS noise and 1/f noise) and ADC remains
Solution Approach 1:
The patent enables the system to correct its own noise by using the visible pixels themselves as part of the correction process. The correction mechanism uses the measured values from masked pixels (which include the same noise characteristics) to correct the visible pixels, allowing the system to self-correct without requiring separate correction circuits for each noise source.
Data Source
AI summary
A method and system is for limiting the x-droop effect in the digital image captured with solid state image sensors with a correction mechanism which instead of using only correction values from the same column to which the correction is applied, also takes the neighboring pixels into account to provide an averaged value to aid in the reduction of temporal and fixed noise contributions associated with the readout of a single pixel.