Nonlinear Analog Gain Controller for Image Sensor Dynamic Range
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
CMOS image sensors face data loss and image quality deterioration due to linear response to light and excessive fixed pattern noise when trying to maintain constant luminosity across varying illumination conditions, especially when the scene's dynamic range equals the sensor's dynamic range.
Innovation Solution
An image sensor with a non-linear analog gain controller and a ramp signal generator that adjusts the slope of the output ramp signal based on illumination conditions, using a non-linear analog gain controller to generate folding currents for voltages higher and lower than the median voltage, allowing for optimal dynamic range management without data loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If a linear response sensor is used with constant ramp signal to maintain constant luminosity, then luminosity stability is improved, but data loss occurs in highlights when analog gain is increased to brighten shadows
Solution Approach 1:
The patent applies local quality by implementing different ramp signal slopes for different illumination regions: a first slope for low illumination portions and a second slope for high illumination portions. This allows the system to brighten shadows without saturating highlights, resolving the contradiction between luminosity stability and data loss in highlights
Solution Approach 2:
The patent implements dynamics by making the ramp signal slope variable rather than constant. The ramp signal generator dynamically adjusts the slope based on illumination conditions, using a steeper slope for dark regions to brighten them and a gentler slope for bright regions to prevent saturation, thus maintaining both luminosity stability and preventing data loss
2Adaptability or versatility
If non-linear analog pixels are introduced to increase dynamic range, then dynamic range is improved, but excessive fixed pattern noise occurs due to transistor mismatch
Solution Approach 1:
The patent introduces an intermediary element - the non-linear analog gain controller - that operates between the linear pixel array and the ADC. This controller applies non-linear gain adjustment to the analog signals after pixel readout but before digital conversion, achieving extended dynamic range processing while avoiding the transistor mismatch issues that would occur if non-linear pixels were implemented directly
Solution Approach 2:
The patent substitutes the physical non-linear pixel structure with a non-linear processing mechanism in the analog domain. Instead of changing the pixel hardware to non-linear, the system uses analog gain control with different slopes to achieve non-linear response characteristics, thereby avoiding the fixed pattern noise inherent in non-linear pixel transistors
3Adaptability or versatility
If log sensor is used to increase dynamic range, then dynamic range is improved, but noise sensitivity increases and image quality deteriorates due to inability to use CDS method
Solution Approach 1:
The patent segments the dynamic range processing into two distinct parts: linear pixel readout with CDS for noise reduction, followed by non-linear analog gain control. This segmentation allows the system to benefit from both the noise reduction capabilities of CDS and the extended dynamic range of non-linear processing, avoiding the noise sensitivity of continuous log sensors
Solution Approach 2:
The patent maintains continuity of useful action by preserving the CDS method's noise reduction function while adding non-linear gain control. The system continuously operates with CDS for reset signal subtraction and then applies non-linear ramp signals for dynamic range extension, ensuring both noise reduction and dynamic range improvement occur simultaneously without interruption
4Adaptability or versatility
If pixel array with non-linear analog pixels is implemented, then dynamic range is improved, but pixel size becomes large due to number of transistors
Solution Approach 1:
The patent extracts the non-linear processing function from the pixel array itself and relocates it to a separate analog gain control circuit. This extraction allows the pixel array to remain compact with standard linear pixel structures, while the non-linear dynamic range processing is performed externally by the ramp signal generator and analog gain controller
Solution Approach 2:
The patent merges the non-linear gain control functionality with the existing analog-to-digital conversion pipeline. The non-linear ramp signal generation and analog gain control are integrated into the signal path between the pixel array and ADC, combining multiple functions (linear readout, CDS, non-linear gain control) into a unified processing chain without increasing pixel area
Data Source
AI summary
An image sensor comprises an active pixel sensor (APS) array, a first analog-to-digital converter (ADC), and a ramp signal generator. The APS array has includes a plurality of pixels of arranged in a second order two-dimensional matrix, and wherein the APS array generates a reset signal and an image signal for each pixel of selected columns. The first ADC has includes correlated double sampling (CDS) circuits for each column of the APS array, and wherein the first ADC generates a digital code corresponding to the difference between the reset signal and the image signal using an output ramp signal that is applied to the CDS circuits for each column. The ramp generator generates the output ramp signal in which a low illumination portion and a high illumination portion have different slopes.