CMOS Image Sensor Correlated Double Sampling Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional CMOS image sensors with global shutters suffer from noise due to uncorrelated read-outs of image and reset signals, leading to decreased signal-to-noise ratio (SNR) when capturing high-speed moving objects.
Innovation Solution
The proposed imaging apparatus and method incorporate a pixel array with capacitors for storing reset and pixel signals, allowing for correlated double sampling by reading and storing these signals at different times, ensuring full correlation and noise elimination through a column parallel reading structure and switch control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If a global shutter is used to capture high-speed moving objects without distortion, then image distortion is eliminated, but noise increases due to uncorrelated read-outs of image and reset signals
Solution Approach 1:
The patent applies preliminary action by reading and storing the reset signal before reading the image signal. The reset signal is captured and held in a storage capacitor in advance, establishing a time-correlated reference that will be subtracted from the image signal to eliminate noise. This preliminary capture of the reset state enables subsequent noise cancellation through correlated double sampling.
Solution Approach 2:
The patent implements feedback through the correlated double sampling process where the reset signal is subtracted from the image signal. The stored reset signal serves as a reference that feeds back into the signal processing chain, allowing the system to actively cancel out noise components present in both signals. This feedback mechanism continuously refines the output by removing correlated noise.
2Object-affected harmful factors
If traditional correlated double sampling is used to reduce noise, then noise is partially eliminated, but full correlation between reset and pixel signals cannot be achieved
Solution Approach 1:
The patent merges the read-out paths of the reset signal and image signal into a unified correlated sampling system. Both signals are read through the same column circuit and processed together in the same storage and subtraction pipeline. This merging ensures that both signals experience identical noise conditions and processing characteristics, achieving full correlation that maximizes noise cancellation effectiveness.
Solution Approach 2:
The patent introduces a storage capacitor as an intermediary element that holds the reset signal while maintaining its temporal correlation with the subsequent image signal. This intermediary storage mechanism allows the reset signal to be preserved in its original state without degradation, serving as a perfect reference for later subtraction. The storage capacitor acts as a buffer that maintains the correlation relationship between the two signals throughout the processing sequence.
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 enables the capture of high-SNR images by ensuring full correlation between reset and pixel signals, reducing noise and distortion, and allowing for high-speed object imaging without the Jell-O effect.
Implementation Method 1
at least one of the pixels comprises a capacitor configured to store a reset signal and a pixel signal
Data Source
AI summary
The present invention relates to an imaging apparatus, including a pixel array comprising a plurality of pixels arranged in rows and columns, wherein at least one of the pixels comprises a capacitor configured to store a reset signal and a pixel signal; and a plurality of column circuits, wherein at least one of the column circuits reads the reset signal and the pixel signal from the capacitor respectively, and generates the difference between the reset signal and the pixel signal, the generation of the pixel signal being later than that of the reset signal.