CMOS Image Sensor Row-Wise Noise Suppression via Analog Capacitive Coupling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional CMOS image sensors face challenges in suppressing row-wise noise in the analog domain, which leads to undesirable horizontal stripes in images due to correlated noise, and existing methods that estimate and subtract noise in the digital domain increase processing overhead and power consumption.
Innovation Solution
The proposed solution involves coupling dark reference pixels to the input of a VCL buffer via capacitors, allowing for analog subtraction of noise from active pixels, thereby suppressing row-wise noise before analog-to-digital conversion, independent of the number of dark reference pixels used, and eliminating noise from the VCL buffer amplifier.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If digital domain noise estimation and subtraction is used, then row-wise noise suppression is achieved, but processing overhead and power consumption increase
Solution Approach 1:
The patent replaces digital domain processing with analog domain processing by coupling dark reference pixels to the input of the VCL buffer via capacitors. This substitution of the processing domain (from digital to analog) eliminates the need for complex digital estimation and subtraction operations, thereby reducing power consumption while maintaining noise suppression effectiveness.
Solution Approach 2:
The patent performs noise suppression in advance by coupling dark reference pixels to the VCL buffer input before the analog-to-digital conversion process. The capacitors store the noise signal from dark reference pixels, which is then subtracted from active pixel signals in the analog domain before conversion, eliminating the need for subsequent digital processing and reducing overall power consumption.
2Object-affected harmful factors
If digital domain noise estimation and subtraction is used, then row-wise noise suppression is achieved, but processing time increases
Solution Approach 1:
The patent replaces computationally intensive digital estimation and subtraction operations with simple analog circuit operations. The capacitive coupling and analog subtraction occur naturally during the readout process, eliminating the need for separate digital processing steps and significantly reducing processing time.
Solution Approach 2:
The patent integrates noise suppression into the continuous analog readout process rather than requiring separate digital processing steps. The capacitors continuously track the noise signal from dark reference pixels, and the analog subtraction occurs continuously during signal readout, maintaining uninterrupted useful action without additional processing delays.
3Measurement precision
If more dark reference pixels are used for digital averaging, then noise estimation accuracy improves, but processing overhead increases
Solution Approach 1:
The patent replaces the complex digital averaging process with a simple analog capacitive coupling mechanism. The capacitors naturally average the noise signals from multiple dark reference pixels in the analog domain through parallel coupling, eliminating the need for digital counting and averaging operations while maintaining accuracy.
Solution Approach 2:
The patent merges multiple dark reference pixel signals through capacitive coupling at the VCL buffer input. The capacitors combine the noise signals from multiple dark reference pixels in parallel, naturally achieving averaging without requiring digital processing overhead for counting and computing averages.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces row-wise noise effectively in the analog domain, minimizing processing overhead and power consumption while maintaining accurate noise correction, and allows for independent use of dark reference pixels across color planes.
Implementation Method 1
CMOS image sensors sense light by taking advantage of photoelectric effect. Photoelectric effect occurs when photons interact with crystallized silicon to energize and move electrons from the valence band into the conduction band, where the electrons can be harnessed, creating electric current at a voltage related to the bandgap energy.
Implementation Method 2
Each light-sensing element includes a photosensor, which may be a photodiode, a photoconductor, a photogate or the like. When a photosensor is exposed to light, it records the intensity or brightness of the light that falls on it by accumulating an electric charge that is proportional to the brightness of the light due to the photoelectric effect.
Data Source
AI summary
Embodiments of circuits and methods for suppressing row-wise noise in the analog domain in an image sensing device. In one embodiment, a pixel sampling circuit includes a readout circuit that is connected to a plurality of pixels to receive analog signals from the pixels. The pixel sampling circuit also includes a noise correction circuit that provides a reference signal to remove at least a portion of the noise in the analog signals received from the pixels before the analog signals are converted into digital signals.